Assignment 5 - Testing Results

Date of submission: February 28th, 2022

Voice-Controlled Wheelchair

Team 13

Pho V Le

Waverly Hampton III

Erik Coleman

Shazib Naveed

Instructors: Russ Tatro, James Cottle

Table of Contents

EXECUTIVE SUMMARYA. Societal Problem

B. Design Idea

C. Work Breakdown StructureD. Project Timeline

ABSTRACT

A. Societal Problem

B. Design Idea

C. Work Breakdown StructureD. Project TimelineE. Risk Assessment

I. INTRODUCTION

A. Societal Problem………………………….…………………..……...…….…..1

B. Design Idea………………………….……..……………….....……………. ...1

C. Work Breakdown Structure……………………………………….………..…..2D. Project Timeline……………………………………………………………......2E. Risk Assessment…………………………………………………………….....2

II. SOCIETAL PROBLEM……………………………………………...……………...3

III. DESIGN IDEAA. Design Philosophy…………………………………...……………8B. Specific Design Components…………………………………..11

PUNCH LIST……………………………………………………………..…….14

IV. FUNDING.………………………………………………………………….15V. WORK BREAKDOWN STRUCTURE ………………….………………16VI. PROJECT MILESTONES AND TIMELINE….…………………….…16VII. RISK ASSESSMENT……………………………………………………VIII. DEPLOYABLE PROTOTYPE STATUS………………………………IX. MARKETABILITY FORECAST………………………………………..

X. CONCLUSION………………………………………………………………….38

APPENDIX D Work Breakdown Structure………………………………….……………24

APPENDIX E Timeline Charts and PERT Diagrams…………………………………….32

APPENDIX F. RESUMES………………………………...…………………………….37

i

Table of Figures

Figure Title Reference

Figure 1 Percentage of Adults w/Disabilities

Richard C. Simpson, PhD,ATP; Edmund F. LoPresti,PhD; Rory A. Cooper, PhD,Department of RehabilitationSciences and Technology;University of Pittsburgh, PA,JRRD Vol. 45 pgs. 53-72

Figure 2 Different Types of People w/Disabilities

https://www.cdc.gov/ncbddd/disabilityandhealth/infographic-disability-impacts-all.html

Figure 3 Patience Wheelchair https://www.grayingwithgrace.com/wheelchairs-narrow-doorways/

Figure 4 Standing Wheelchair https://livingspinal.com/active-mobility/standing-wheelchairs-and-frames/draco-electric-standing-wheelchair/

Figure 5 Sports Wheelchair https://magazine.viterbi.usc.edu/spring-2017/alumni/designing-the-worlds-fastest-wheelchair/

Figure 6 Flowchart of Control SystemDesign

Erik Coleman

Figure 7 Dual G2 High-Power MotorDriver from Pololu

Pololu

Figure 8 Electric DC Motor Polulu

Figure 9 Flowchart of program Erik Coleman

Figure 10 Lidar function code Shazib and Erik

Figure 11 State control function Shazib and Erik

Figure 12 Motor control function Shazib and Erik

Figure 13 Risk Matrix Erik and Waverly

ii

Table of Tables

Table Title Reference

Table 1 Punch List Waverly Hampton III &Erik Coleman

Table 2 Funding Table Waverly Hampton III

Table 3 Work Breakdown Structure Team

iii

Executive SummaryA. Societal Problem

The year is 2022 and access tomobility is a problem we all face.However, there are those who have lessaccess than others. People with disabilitiesare limited in their range of motion, andworst of all, the products they rely on arelimited as well. The next generation ofmobility for those with disabilities startswith the wheelchair. A device that has, fordecades, not been improved upon much,the wheelchair is due for an upgrade sothat those with limited or no ability tomove can be in motion once again.

B. Design IdeaTo change the wheelchair, one must

first change the way one thinks about whata wheelchair’s function is when used. Awheelchair must move a load from point ato point b, assisted by some form of power.Our design is the next generation of thewheelchair. It not only uses voicecommands to input directions and electricpower to drive the motors, it is alsoequipped with sensors that give it theability to detect objects around it and avoidcollisions. Safety is the primary feature ofour design, for our wheelchair will be ableto measure its own speed and adjust itsposition so that it never puts the user in adangerous position.

C. Work Breakdown StructureThis project will consist of a

variety of tasks to complete, including butnot limited to: Designing and building theelectric drivetrain; programming theRaspberry Pi module for interfacing with

the voice control module and sensors;programming the motor driver to controlwheel speed and direction; demoingprototype, and revising its design;completing a market analysis; and lastlydelivering the project on time.

D. Project TimelineThe project’s timeline allows us to

track our progress throughout the semester.During the second semester, we did changethe timeline due to some parts of theproject being pushed back. We had issueswith movement during the first semester,so our timeline for some of the workpackages for movement was pushed to thesecond semester. The project timelineallowed us to keep track of our workpackages and make sure that we did notfall behind.

E. Risk AssessmentThe Risk Assessment allows us to

predict any risks that might occurthroughout the semester. For the riskassessment we made a Risk Matrix thatoutlines the risks. The risks are then givena probability. This allows us to predict ifthe risks might happen and gives us theopportunity to create back up plans in casethe risks occur.

G. Device Test PlanTest Plan will be a beneficial tool

to keep us on track of the project timeline.Since the project has lots of tasks that needto be completed, some functions will bechanged overtime. A testing plan willmake sure we give us an idea of how manyhours we should spend on each task. Alsohelp us predict what problems can happen

during time. This senior project is2-semester long. Our test plan in thissecond semester was changed so it iscloser to the requirement and ourexpectation.

H. Market ReviewThe Market Review allows us to

see other types of products that are offeredfor disabled people. There are variousdifferent mobility products for disabledpeople. For example, there arewheelchairs, electric wheelchairs, andother various tools that help the person thathas difficulty moving. Looking at ourmarket allowed us to understand thevarious different tools that are present to adisabled person and the differencesbetween our wheelchairs versus the otherproducts on the market.

I. Testing ResultsThe mission of team 13 project is

to create a voice-controlled wheelchairwhich can improve the moving ability ofpeople with disabilities. In this section oftesting, we focused on our final result afterthe wheelchair is completely assembledwith all components. There were somemodifications made at the end to adapt toour current situation.

ABSTRACTA. Societal Problem

We can see that the world is in aperiod of constant change, becoming morecivilized and modern. The more modernlife is, the more it cannot be without thepresence of electronic devices. Thesedevices appear everywhere to serve theinterests of people, from daily life toproduction. Consumer electronic devicesfocus on accuracy and speed. And amongthem, the technology that was beingdeveloped and favored at that time wasremote control technology. It has made agreat contribution towards controllingdevices for people who need to sit in placewithout physically operating equipment.Currently, the largest application for thiselectronic remote-control industry isintegration with the design of smart homes,which is quite popular today.

B. Design IdeaWe want to apply our programming

knowledge to implement a meaningfuldesign. That's why we chose the topic"Controlling an electric wheelchair byvoice or smartphone" as our senior project.This is a meaningful topic for humanity,which will assist disabled people who areunable to control wheelchairs with theirhands or feet.

C. Work Breakdown StructureThis project’s timeline is from

August 2021 to May 2021. Tasks includebut are not limited to: coming up with asocietal problem, designing a solution tothe societal problem, creating a workbreakdown structure to map the project,developing a prototype to demonstrate,

redesigning prototype for any refinements,completing a market analysis, anddelivering the project on time. The finalprototype should be completed by the endof April.

D. Project TimelineThis project has several

components that need to communicatewith each other. To successfully completethis design, we came up with a timelinewhich describes each work package. Thistimeline will allow us to track our progressthroughout the semester. This also allowsus to see dependencies betweencomponents and ensures we notice anybottlenecks in our design approach.

E. Risk AssessmentThis assignment attempts to

identify any foreseeable risks with ourproject. This will allow us to be preparedto mitigate any issues that do arise. To stayon track, we need to be able to moveforward no matter the issue. Identifyingand describing any possible issues willhelp us do so.

G. Device Test PlanTesting is one of the most

important steps in this whole designingproject. After completion with each featurewe will execute some test for each part tomake sure it is working properly beforeinstallation. The tests will be performedfrequently before and after each demo.This will give us ideas of what needs to beimproved or estimation of time theprototype would be completed. Also keepus on track in case there are any problemsidentified.

H. Market ReviewMarket research includes a number

of analytical activities on: marketcompetition, market development trends,market segments, buyer behavior andpsychology, effective distribution channelsand the benefit that the product willprovide for customers. With the diversityin the wheelchair market, people withdisabilities have a lot of good choiceswhich serve their needs appropriately.Depending on the status of their disability,customers will make their selection. Thisproject we are targeting the customer whocan’t control their wheelchair with bothhands and feet.

I. Testing ResultsTesting is the most important step

before we can complete our design. Inreality, the testing is supposed to beexecuted throughout each task when theyare completed. When all the individualfeatures were tested successfully, they willbe assembled. At this point the final testingis executed. The testing needs to berepeated multiple times to ensure thedesired results are achieved. Anddepending on the result of each test, a newmodification will be applied until the resultis satisfied.

I. IntroductionA. Societal Problem

The world is in the wave of theindustrial revolution 4.0. The technologyof intelligent control and automatic controlalso develops, they are applied in manyfields in industry, life and even medicine.According to Patric R. Spence, “Robotsand other machine communicators areincreasingly performing social andworkplace roles such as teachers,caregivers, surveillance, decision-makersand personal companionship’’. Recently,when people's living needs are improvingday by day, these technologies are gettingmore attention. Speech recognition is atopic that has received the attention andresearch of many scientists over the pastdecades. Speech and writing are the twomost basic components of language,speech formed before writing anddeveloped throughout the history of humandevelopment. Voice is and will always bethe primary tool of human communicationbecause communication by voice issimple, natural and plays a key role inhuman life. Nowadays, along with thedevelopment of science and technology,scientists have created machines thatgradually replace manual labor and thehuman brain, but human-machinecommunication is still popular. Theoperations are complex operations andrequire training. For that reason, theresearch and development of devicescapable of communicating with peoplethrough voice has been receiving theattention of many scientists around the

world, as well as consumers, but the levelof popularity is still limited.Biometrics - or technology that usesbiological characteristics of people toidentify is a very diverse field and hasmany important practical applications. Inthe field of biometrics, speech has receiveda lot of attention due to the naturalness ofthe voice, the ease of collecting and usingthe voice in the process of recognizing thespeaker. Many methods have been studiedand achieved certain effects in the processof speaker identification. Human speechrecognition has been attracting the researchattention of many scientists as automationtechnology has more and moreapplications in real life.The goal of this topic is to use the voicerecognition in the circuit and combine itwith our smartphone control to apply it towheelchair control. The wheelchair will becontrolled by the user to be able to goforward, backward, turn left, turn right andaccelerate and decelerate speed accordingto needs.

B. Design IdeaTo accomplish the goal of creating the nextgeneration of wheelchair, we decided toimplement a variety of sensors, includingLidar, motion and a forward facing RGBcamera to allow the wheelchair to be awareof its surroundings. The idea is that thenext generation of wheelchairs will beresponsive to the environment to create asafer experience for the user. Safety is themain feature of our design, because thosewith disabilities are already burdenedenough, so our responsibility as engineers

is to design a product that is ergonomicand makes their lives easier.

C. Work Breakdown StructureThe criteria for determining te22he

status and expected delivery of this projectwill be based on breaking down the projectfeatures, listed in the design approach, intodoable tasks, subtasks, and work packages.As shown in Table 3, located in AppendixD, the tasks will be defined as the Feature,the subtasks will be thought of asRequirements, and the work packages arethe various assignments needed to becompleted to meet each requirement. Eachassignment has an anticipated due date,making it possible to chart a schedule forthis project.

D. Project TimelineWe outlined our project timeline using

a Gantt chart in Microsoft excel. We brokeup our tasks to align with our workbreakdown structure. Each feature is listedunder its own section, along with workpackages assigned. The courseassignments for both semesters are listedunder section 7 & 8, for Fall and Springsemester.

E. Risk Assessment

The voice-controlled wheelchair isprone to various risks. Most importantly,the wheelchair is powered by electricityand has a standby battery, prone to failureleading to inconveniences. The wheelchairdepends on voice commands from the user,which could be contradicting if the user is

in a noisy environment. Again, thewheelchair is not able to foresee dangerahead; this puts the user at risk ofencountering dangers that might leavethem hurt. In every street, some evil peopleare never passionate, and this being a merechair puts the user at risk of being attackedby robbers or other evil-minded persons.

A broader technical risk related toour sensor system, is its ability to detectobjects in our path. We are concerned withstationary and moving objects that couldbe in our intended path. We will need totest the Lidar’s ability to detect thesepossibilities and determine if it will besufficient to meet our intended designmetrics.

G. Device Test PlanWe have devised several tests for

our prototype. Each design element has atest to confirm its operation by itself.. Eachfeature has a series of tests, which willallow us to confirm the correct operationof our system. We have defined the testswith our environment and our own biasesin mind. Once we confirm each part of oursystem is functioning as we expect, wehave functional tests to observe theprototype as a whole unit. These tests willdemonstrate our systems ability to performbased on our defined measurable metrics.

H. Market ReviewWith the variety of wheelchairs in

the market, our voice-controlled wheelchair will be the unique product whichsupports disable patients. Wheelchair

market with the diversity of price is aproblem which we have to solve. Thedifference in price of technologywheelchair and traditional wheelchair is abarrier in the process of approachingcustomers. Market review is an importantstep in the whole project timeline

I. Testing ResultsAfter about 32 weeks of

researching and executing the project,team 13 has achieved the initial objectiveswhich were suggested at the first semester.We have designed the wheelchair systemwith voice controls which can control thewheelchair to go forward, reserve , turnleft, turn right , change speed and avoidcollision. This is the topic that the teamonly did on an experimental basis and hasnot been applied in practice. So for thismodel to be more completed, there will besome options which we propose in theconclusion to improve the model and canbe applied in practice.

II. Societal ProblemA. First Semester Interpretation of the

Societal ProblemThere are different types of

disabilities that humans can develop.These different types of disabilities eachhave their own problems that the persontends to deal with daily. Humans developmobility limitations when they are born orthrough an injury that leaves them unableto walk. “Mobility limitations are theleading cause of functional limitationsamong adults with an estimated prevalenceof 40 per 1,000 persons aged 18 to 44 and188 per 1,000 aged 85 years and older inthe general population” [14].

Figure 1 – Percentage of adults that have differentdisabilities [14].

Figure 1 shows the different types ofdisabilities that adults have. In the figure,mobility disabilities are the highestpercentage that adults have. 13.7% ofadults have mobility disabilities. Mobilitydisability requires a person to have sometype of walking support for a person to getfrom point a to point b.

Alzheimer's disease is a type ofdisease in which people with the disease

tend to use a wheelchair. People withAlzheimer's tend to lose functions such aslanguage, motor skills, and perception.This disease causes people to developmobility disability over time. As time goeson it becomes harder for them to get frompoint a to point b, as a result this alwaysrequires another person to help wheel themin the wheelchair from point a to point b.

Amyotrophic Lateral Sclerosis is aneuron disease which causes people todegenerate motor neurons over time. Astime goes on, this leads to weakness inpretty much the “The muscle weaknessassociated with ALS can interfere with anindividual's ability to operate a wheel-chairin two ways: (1) weakness or fatigue in thearms and shoulders can prevent personsfrom completing long routes and (2)weakness in the neck muscles can makeseeing obstacles behind or next to thewheelchair difficult or impossible …Every person diagnosed with ALS willeventually need a wheelchair” [14].

Cerebral Palsy is a brain injuryoccurring before cerebral development iscomplete. Cerebral Palsy can result inphysical dysfunction. “Of individuals withCP, 86 percent use a wheelchair at leastsome of the time” [14]. This injuryrequires the person to use the wheelchairwhen traveling long distances.

Multiple sclerosis damages themyelin sheath surrounding nerve fibers inthe central nervous system [14]. Symptomsinclude fatigue, tremors, and impairedreasoning. “Half of individuals with MSrequire the assistance of another person foreveryday mobility … Of individuals withMS who use manual wheelchairs, 59percent stated they did not feel theircurrent wheelchair met their mobility”[14]. Many of these injuries and diseasescause people to lose mobility with awheelchair being their only option to movearound.

Spinal Cord Injury at or abovefourth Cervical Vertebra can cause peopleto not be able to operate a manualwheelchair. Their lack of neck movementmakes it impossible for them to seeobstacles behind or next to them. “SCI atany level has been estimated to beprevalent in 200,000 to 288,000 peopleliving in the United States with an SCI”[14].

Figure 2 – Different types of people with disabilitiesbased on race, gender, and age [15].

The figure above demonstrates that40% percent of adults that are 65 years andolder have a disability. It shows that 1 in 4women have a disability, and that 2 in 5 for

non-Hispanic American Indians/ AlaskaNative have disabilities. This proves thatthere is a huge need for consumer productsthat provide aid to these marginalizedgroups. One of the most popular of theseproducts is the wheelchair.

A wheelchair is a chair with wheelsthat is used when walking is difficult orimpossible due to illness, injury, ordisability. Wheelchairs come in a varietyof formats to meet the specific needs of theuser. These can include specialized seatingadaptations, individual controls, and can bespecific to specific activities, as seen withsport wheelchairs and beach wheelchairs.The most widely recognized distinction isbetween powered wheelchairs ('electricvehicles'), in which propulsion is providedby batteries and an electric motor, andmanual wheelchairs, in which propulsionis provided. provided by the user, thewheelchair user pushing the wheelchairmanually ('self-propelled') or pushed fromthe back by an attendant ('steward push').

The earliest records of wheeledfurniture are an inscription found on astone slab in China and a children's beddepicted in a Greek vase painting, bothdating from 6th to 5th centuries BC. Thefirst records of wheeled chairs used totransport people with disabilities date backto three centuries later in China; TheChinese used wheelbarrows early to movepeople as well as heavy objects. Adistinction between the two functions wasnot realized for another few hundred years,until around AD 525, when images ofchairs with wheels were introduced, and

specially designed chairs made to carrypeople began to appear in Chinese art.

Although the Europeans eventuallydeveloped a similar design, this method oftransportation did not exist until 1595when an unknown inventor from Spainbuilt one for King Phillip II. Although itwas a complex chair with both armrestsand footrests, the design was flawed as itdid not have an effective push mechanismand therefore required support to push it.This makes the design more like a modernor portable armchair for the rich peoplethan a modern wheelchair for the disabled.

In 1655, Stephan Farffler, a22-year-old Paraplegic watchmaker, builtthe world's first self-propelled chair on athree-wheel chassis using a shaft and gearsystem. However, the device has the shapeof a hand bike rather than a wheelchairbecause the design includes cranksmounted on the front wheel.

In 1887, wheelchairs wereintroduced to Atlantic City for tourists.You can rent them to enjoy the Boardwalk.Soon, many healthy tourists were alsorenting decorated "wheelchairs" andservants to push them as a display ofdecadence and treatment they could neverexperience.

In 1933 Harry C. Jennings, Sr. andhis disabled friend Herbert Everest, bothmechanical engineers, invented the firstlightweight, steel, folding, portablewheelchair. Everest had previously brokenhis back in a mining accident. Everest andJennings saw the business potential of theinvention and went on to become the firstmass-market wheelchair manufacturer.

Over the years since then, a variety ofwheelchair types have been introduced.

First, the Manual Wheelchair: Amanual wheelchair combines a frame, seat,one or two legs (legroom) and four wheels:usually two wheels in the front and twolarge wheels in the back. There will alsogenerally be a separate seat cushion.Larger rear wheels usually have a push rimof a slightly smaller diameter, extendingbeyond the tire; these allow the user tocontrol the seats by pushing on themwithout requiring them to grip the tire.Manual wheelchairs usually haveload-bearing brakes on the tires of the rearwheels; however, these are only parkingand motion brakes provided by the user'spalm directly on the pusher rim. As thiscauses friction and heat build-up,especially on long descents, manywheelchair users will choose to wearwheelchair gloves.

Figure 3: Patience Wheelchair[6]

Second, the Patience Wheelchair:A patience wheelchair is similar to amanual wheelchair, yet it has smalldiameter wheels on both front and rear.

The chair is driven and controlled by aperson standing in the back and pushingthe handle integrated into the frame.Brakes are provided directly by the valet,who will usually be provided with a foot ormanual parking brake.

Third, the Power Wheelchair: Apower wheelchair, commonly referred toas an “electric wheelchair", is a wheelchairthat incorporates a battery and an electricmotor into the frame and is controlled bythe user or attendant, most commonly via asmall joystick, mounted on the armrest oron the upper rear of the frame. For userswho cannot manage manual joysticks,head switches, chin joysticks, sip-and-puffcontrols, or other expert controls that canenable independent operation of thewheelchair. Ranges of more than 10 miles,15 kilometers are generally available fromstandard batteries. Powerchairs are oftendivided according to their accessibility. Anindoor chair can only reliably pass overperfectly flat surfaces, limiting them tohousehold use. An indoor-outdoor chair isless restrictive, but may have a limitedrange or ability to cope with slopes orsurfaces not flat. An outdoor chair is morecapable, but will still have a very limitedability to deal with rough terrain. A veryfew specialist designers offer a truecross-country capability.

Figure 4: Standing Wheelchair [8]Fourth, the Standing Wheelchair: A

standing wheelchair is one that supportsthe user in an almost standing position.They can be used as both a wheelchair anda standing frame, allowing the user to sit orstand in the wheelchair at will. Someversions arfully manual, others have amanual seat stand, while others have fullpower, tilt, recline and variations of thestand function available. The benefits ofsuch a device include, but are not limitedto: supporting independence andproductivity, enhancing self-esteem andpsychological well-being, enhancing socialstatus, expanding access, reducingpressure, relieve pressure ulcers, improvefunction, improve respiratory function,reduce occurrence of UTIs, improveflexibility, help maintain bone mineraldensity, improve passive range motion,

reduce Abnormal muscle tone andspasticity, and bone deformities. Otherwheelchairs provide some of the samebenefits by raising the entire seat to raisethe user to standing height

Figure 5: Sports Wheelchair [9]Last, but not least, the Sports

Wheelchair: The sports wheelchair is usedin competition. A variety of disabilitysports have been developed for athleteswith disabilities, including basketball,rugby, tennis, racing and dance.Wheelchairs used for each sport haveevolved to fit the specific needs of thatsport and are often no longer the same astheir everyday cousins. They usually donot fold (to increase stiffness), providingstability and useful for making sharp turns,and are usually made of lightweight,composite materials. Even the seatingposition can be completely different, withracing wheelchairs often used in a kneelingposition. Sports wheelchairs are rarelysuitable for daily use and are often a

“second” seat exclusively for sports use,although some users prefer sport optionsfor everyday use. Some people withdisabilities, specifically the lower limbs,can use wheelchairs for sports, but not foreveryday activities.Each of these different wheelchairs haveone thing in common: They all requiresome form of manual input from the user.For those with severe disabilities andlimited range of motion or use of theirhands, a modern wheelchair is notadvanced enough. We strive to solve thisproblem by asking the question: In today’sworld of Roombas and robots, why is itthat wheelchairs still require the user to lifta finger?

B. Second Semester Interpretation ofthe Societal Problem

After testing our initial design, we realizedthere were a number of things we shouldhave done differently. Our prototype useda L289n motor driver to supply power toour DC brushed motors. While testing, wedetermined the continuous current outputfrom the driver was not sufficient toovercome the stall current of the motors.This caused our motors to stall if there wasany weight set on them.

We then bought a high powermotor driver from Pololu. Its twin discreteMOSFET H-bridges support a wide 6.5 Vto 36 V operating range with a continuouscurrent of 18A. This should be more thanenough power for our design. After doingthe initial setup and testing, we confirmedthe motor driver was operating correctly.Our prototype was able to traverse acrossthe room, with one caveat. The weight ofour 12V lead acid battery was too muchfor the motors to handle.

Our research has shown that ourmotors do not have enough torque as is.Drive systems typically take advantage ofgears to provide higher torque to thewheels. Having enough torque is criticalfor our design, in order for a person to rideon the wheelchair.

III. Design IdeaA. Design Philosophy

Our design will allow disabledpeople to navigate an electric wheelchair,without the need for manual control. Wewill be giving the user the ability to controlthe wheelchair using simple voicecommands. To ensure the safety of theuser, we have a sensor to allow thewheelchair to traverse through ahospital-like setting and avoid objectcollision.. The core functionality of ourdesign will be mobility, voice control, andobject detection.

It was always the intention for theelectric wheelchair to be battery-poweredand chargeable. The main operatingcontrol system consists of amicroprocessor, a microphone, speedcontrollers, LiDar sensor, and two DCmotors. The LiDar sensor is fixed to thefront of the wheelchair, to determine thewheelchair’s distance from an object in itspath. This is necessary since the DCbrushed motor does not have speedcontrol. Magnets will be fixed equidistantalong the circumference of the wheel,which will enable the Hall effect sensor totrigger a signal to the Raspberry PI. Inbetween signals a timer will be counting,which will help us determine how fast weare moving. We will be using a pulse widthmodulated signal to control our speed anddirection of the DC motors. Based on ourcurrent position and command, we willvary the voltage sent to the H-bridge, tocontrol the direction our wheelchair willmove.

We are going to have a microphoneconnected to the Raspberry Pi, which willrecord the user’s speech in real time. Wethen will use speech recognition todetermine the intended movement from theuser. To utilize Google’s speech API, theraspberry pi will need to have a constantinternet connection. We will need todetermine if this will fit our use-case, if notanother speech recognition method willneed to be implemented. We intend to haveour wheelchair accept numeroususer-friendly commands for operation.When the user intends to moveforward,commands such as ‘go,’’forward,’ ’straight’ should all cause thewheelchair to move in a forward fashion.If the user wants to maneuver around aturn, they will need to give a stopcommand and then a command to turneither left or right. If a left turn isoccurring, the raspberry pi would send apositive voltage to the right wheel and anegative voltage to the left wheel. Thiswill allow our wheelchair to have a smallerturn radius.

Figure 6 - Flowchart of control system design. Designedusing app.diagrams.net

Commands will not be executed ifsomething is blocking the path, or a fallwould potentially occur. These scenarioswill be detected by either our camera orlidar sensors. The wheelchair would thenbe put into an idle state, waiting for inputfrom the user. This will prevent anaccident from occurring. The primaryconcern of this design is the safety of ouruser, so this will need to be testedthoroughly. Another reason for a commandto potentially be blocked is if the voice isnot recognized. Malintent from anotherperson is a potential cause of disaster. Ifour wheelchair could easily takecommands from anyone within speakingdistance, this would allow our system tobehave in unintended ways. We will needto verify the command is in fact beingissued by the owner of the system.

Sensors on our system will allowus to react to unforeseen circumstancesthat may arise. We plan to equip thewheelchair with an RGB camera facingforward, to detect any obstruction in ourpath. The control system will then decideto either maneuver around the object orpause and wait for further direction. Twolidar sensors will need to be used, facingdirectly in front and behind the wheelchair.These will be used to determine if a ledgeis about to occur.

Figure 7 - Dual G2 High-PowerMotor Driver from Pololu (24v18 forRaspberry Pi)

The dual motors can be controlledindependently. Each motor has 4 GPIOpins associated with it. A fault pin, PWMinput, enable, and a direction pin. Thesecan be seen in the table below. We areusing the Broadcom chip-specific pinnumbers(BCM).

Value GPIO Pin (BCM)

Motor 1 Fault 5

Motor 2 Fault 6

Motor 1 PWM 12

Motor 2 PWM 13

Motor 1 Enable 22

Motor 2 Enable 23

Motor 1 Direction 24

Motor 2 Direction 25

We are sending a 20kHz hardwarebased pulse width modulated (PWM)signal using the PiGpio Python library. Wecan control the speed of our wheelchair byadjusting the duty cycle of our PWMsignal. The direction pin can be toggled,which will change the direction our motorsspin. ee

Figure 8 – Electric DC Motor (24V)In this version of the project we are

using a built-in gearbox motor with 24Vpower. This type of motor is a DC motorwith 10:1 metal gearbox and integratedquadrature encoder that provides aresolution of 64 counts per revolution ofmotor shaft. With the old version we havefaced challenges of getting the wheelchairto move because of the torque issue. Thepower of 24V in this motor along with abuilt-in gearbox, we believed will solve theproblem of getting the wheelchair to movedue to the heavy load. Another applicationof this new motor is the CPR encoder. Theencoder will send feedback signals that canbe used to determine position, speed anddirection.

Final Design

B. Specific Design Components

Electric Drivetrain: The electricdrivetrain feature will be developed usingfour dc motors, one H-bridge motor driver,and one twelve-volt battery. Pulse WidthModulation will be used to adjust thespeed of the motors. Team member Phowill be responsible for implementing theelectric drivetrain. His background andeducation in electrical engineering will beuseful in designing the proper circuitdiagram and building out the drivetrain.This feature should take an estimated 15hours to complete. For the feature to beconsidered complete the wheelchair’swheels must rotate when a voltage isapplied to the dc motors.

360 Degree Turning: The turningfeature will be developed using a singlewheelchair, its four wheels, four motors,and a voice command. Team memberShazib will be responsible for the task ofimplementing the turning feature. Hisbackground and education in computerengineering will be useful in designing thestate diagram that will be used to programthe wheelchair to turn on command. Thisfeature should take an estimated 20 hoursto complete. For the feature to beconsidered complete the wheelchair musthave a turning radius of 360 degrees.

Collision Avoidance and ObjectDetection: The features of Collisionavoidance and object detection will bedeveloped using one RGB camera, twomotion sensors, and two Lidar sensors.Team member Waverly will be tasked withbuilding out the sensor control system. Hisbackground and education in electrical andelectronics engineering will allow him toapply knowledge of control systems,signals, microcontrollers, and circuitdesign to develop the feature so thewheelchair will be able to detect objectssurrounding it and avoid them. The featureshould take an estimated 48 hours tocomplete. For the feature to be consideredcomplete the wheelchair must detectobjects and then halt or change directionwhen objects are in its path.

Forward and Reverse Accelerationand Speed Detection: The features ofacceleration and speed detection will bedeveloped using Pulse Width Modulationand an accelerometer. For backup, two halleffect sensors with magnets will be utilizedin case the accelerometer is notfunctioning. All team members will beresponsible for developing these features.Our backgrounds and education inelectrical and computer engineering allowus to assemble all the previous featurestogether to enable acceleration and speeddetection to happen. These features wouldtake an estimated 96 hours to complete.The features will be considered completewhen the wheelchair can change positionsand adjust its speed based on the positionof surrounding objects.

Voice Control: The voice control isrunning locally on the Raspberry PI4. Weare running Coqui STT (Speech-to-text),which allows us to stream audio from theUSB microphone, to our transcriber. Oncewe identify a command in the audiostream, our key value is updated. This keyvalue is shared between our multipleprocesses.

The main commands we plan onfeaturing are forward, reverse, turn left,turn right, stop, increase and decreasespeed. Alternative keywords can be used,such as go, start, begin, speed up, faster,slow, slower.

Figure 9 - Flowchart diagram ofPython processes to control motors andsensors.

We choose to use multiprocessingto allow our code to run in parallel acrossall four cores of the Raspberry Pi 4. Themain function creates 3 new processes, andthen begins to stream audio to thetranscriber. Once a command keyword isdetected, the key value is updated. Thisvalue is used to determine which state themotors should be in.

Figure 10 - Lidar function whichimplements object detection. Once a valuebelow the threshold is returned, we set akey value of 8 and start braking.

Figure 11 - State control functionshown below that takes our key value andoutputs a state value to our motor controlfunction.

Figure 12 - This program describesour motor state. With the state value, ourcontinuously running while loopdetermines which state to be in. Once wereach Max speed, this speed is kept until achange of state is issued.

Feature Requirement Measurable Metrics Measurements for SuccessVoice Control Wheelchair must be capable

of moving using input voicecommands

Wheelchair’s positionbefore and after input of

voice commands

Speech recognition can identifycommand keywords and send

the corresponding action to thecontrol until.

Electric Drivetrain Wheelchair's wheels must bepowered by an electric

battery

The rotation of wheelswhen a voltage is applied

The wheelchair must have theability to move in a smooth,

straight fashion.

360 Degree Turning Wheelchair must have aturning radius of 360 degrees

The starting angle ofposition of the

wheelchair’s front beforeand after turning

command is input

The wheelchair should be able tomake a full 360° turn, while

using minimal amount of space.

Collision Avoidance Wheelchair in motion mustnot hit another object

Force applied towheelchair after coming

to a complete stop

The wheelchair must stop oncean object is detected.

Forward andReverse

Acceleration

Wheelchair must be capableof accelerating forward and

in reverse

Wheelchair’s positionbefore and after input ofcommands for movingforward and in reverse

The wheelchair should be able totraverse safely in a forward and

reverse type motion. This shouldbe a smooth motion.

Speed Detection andMeasurement

Wheelchair must be capableof detecting and measuring

its own velocity

Wheelchair’s velocitywith respect to another

object moving at aconstant speed

The system must be able tocalculate the current speed atwhich it is traveling. We can

verify this by measuring speedwith alternate methods.

Object Detection The wheelchair must becapable of sensing objects

around itself

Wheelchair’s response topresence object in front,behind, and on its sides

The sensor system on board canidentify objects in path.

Table 1: Punch List of Feature

IV. FundingTable 2 is a Funding Table listing all items to be purchased and an estimated total cost of$433.00 to complete the project. Total cost will be divided amongst the four teammembers. Each member will be tasked with purchasing and delivering assigned items forfinal assembly. The working budget for this project is $1000.00, therefore this project iscurrently under budget. Remaining funds not spent building prototype will be set asidefor purchasing tools and miscellaneous costs.

Item Quantity Cost EstimatePer Unit

Manufacturer Supplier

Wheelchair 1 1 $110 TBD TBDBrushless DC Motor 2 $35 TBD TBDMicrocontroller(Raspberry Pi 4)

1 1 $70 TBD TBD

Lidar Sensor 2 $15 TBD TBDLiPo Battery 4 $20 TBD TBDVESC 2

Budget $1000 TBD TBDTotal Cost Est. $433 TBD TBD

Table 2: Funding Table

V. Work Breakdown StructureThis project has approximately

thirty-three weeks of activity, frominception of the societal problem to thedelivery of the final design ideaprototype. To successfully approach thedevelopment of the final prototype, thedesign approach has been broken up intofeatures. Those features were thendivided into subtasks based on therequirements needed to complete eachfeature. The subtasks were then divided

into work packages filled withassignments needed to be completed.

VI. Project Timeline andMilestones

In Appendix E, we outline our tasksfor the entire project. Each task isassigned to either an individualteammate, or to the entire team. This canbe changed, if timing or schedulingissues arise. We made sure to giveenough time, to account for any personaltiming issues. Having this timeline willallow us to stay on track and completeeach assignment in a timely manner. Wemade sure to align the project timelinewith our Work Breakdown structure,splitting up the sections by our featureset and assignments.

VII. Risk AssessmentAs described in the introduction, our

design has multiple possible risksassociated with it. These vary fromspecific technical risks to broadertechnical risks as well as systematicrisks.

A broader technical risk isrelated to our wheelchair’s drivetrain.We want the wheelchair to be able tomaneuver around corners with a smoothmotion. Our current design approach isincluding two brushed DC motors, withone on each side. Our design is going tobe two-wheel drive. We plan on varyingthe speed of each motor, to allow thewheelchair to go around corners. It is apossibility that our turn radius will notbe small enough, which would preventus from turning smoothly. To mitigatethis issue, we would add additionalmotors, and make the final design afour-wheel drive vehicle. Allowing us tocontrol each wheel rather than thetwo-wheel drive approach. This wouldadd extra components and increase ourpower requirements. We will need to testthe performance of our design anddetermine if two motors will besufficient.

A specific technical risk isrelated to our chosen microphone’sspeech recognition software andkeyword identification. On the productdescription, it claimed to have offlinekeyword identification and presetcommands available to program with.This functionality seems to be removed

from their current software package. Tomitigate this issue, we will be usinganother approach for speech recognition.This adds time needed to develop asolution for speech control.

A voice-controlled wheelchaircannot foresee dangers ahead. Althoughthe wheelchair is fitted with a Lidarsensor to detect objects, it cannot predictthe movement of obstacles in the path.The wheelchair user will thus have to bevery keen to avoid any risks [16]. Aspecific technical risk is related to theaccuracy of our Garmin Lidar sensor. Wewill need to test the Lidar in certainoperating environments, to determine ifit is able to detect objects in varyinglighting conditions. To mitigate thisissue, we could either add motionsensors, ultrasonic sensors, or a RGBcamera. This will add additionalcomplexity to the system. A potentialbroader technical risk related to theLidar sensor, is having the ability todistinguish between moving andstationary objects. We will need to testthis functionality and see if it issufficient for our purpose. If the machineis unable to tell the difference, our planof action is to add additional sensors,such as a motion sensor on the left andright side of the chair, and possibly anultrasonic sensor.

A broader technical risk is ourability to adjust the speed of ourwheelchair. DC motors have no inherentmethod to determine current speed. Ifour motors are not identical, we mightexperience a stray in our motion over

time. If this is experienced, we will needto add the ability to determine our speed,and adjust using a feedback control loop,such as a PID controller. Sensor datafrom hall effect sensors will determineour current speed and adjust our outputto maintain our desired speed. Thiswould add additional sensors andpossibly a microcontroller to our design.

Confusing voice commandsoccur when the user is in a noisyenvironment, which makes itchallenging to recognize voicecommands. In addition, while in a noisyenvironment, the voice systems arelikely to respond to unwanted voicecommands and head the user in anunexpected direction if no furthercontrol is done [16]. The wheelchair’ssystem motors might misbehave due tohigh input, which speeds the wheelchairbeyond control and may cause accidentsin a busy road.

The wheelchair user is vulnerableto attack by outsiders; - the wheelchair isjust open, which puts the user at risk ofbeing attacked by evil persons they crosson their path. Besides, the user isdisabled and might not fight back theattacker unless somebody comes to theirrescue [17].

A systematic risk associated withdoing teamwork is preventing possibleCovid-19 exposures and practicingpreventative measures. We are followingall guidelines, including wearing maskswhen in close contact. If anyone on ourteam is experiencing symptoms, we

insist they stay home and not come tocampus. This ensures we are notcontributing to any possible spread ofthe virus.

Other possible systematic risksinclude environmental disasters such asfires and flooding. If these were to occur,we would continue to work remotely toensure the safety of our team. Thesehave a low probability of occurring butcould still potentially be an issue.

The impact on the matrix is theamount of money needed to mitigate therisk. This cost is an estimation based onextra components potentially needed.

Figure 13 - Risk matrix showing associated riskvalue with identified risks.

VIII. Deployable Prototype Status

Test ID Test Description Expected Results Actual Results Pass/Fail

1 Check Raspberry PiOK/ACT LED statuson boot. A flashingpattern at bootindicates an error withthe SD-Card.

A consistently litgreen light.

The raspberry piis powering on asexpected and thegreen LED isconsistent.

Pass

2 Check the RaspberryPi Power LED onboot. If the LED isblinking, that indicatesa low power source. Ifit goes out, this isconsidered a“brownout”.

A consistently lit redlight. This indicatesthe RPi is receivingadequate power. Ifthe power supplygoes below 4.63V,the LED will blink.

The RPi shows aconsistent redLED oncepowered on. Weare receivingadequate power.

Pass

3 Run the microphoneV.A.D. streamingpython program onthe RPi 4 and verifykeyword detection.This will verify CoquiS.T.T is installedproperly and themicrophone isfunctioning. Speakinto the microphonefrom 18”-36”. Speakeach keyword 3 timesand verify thetranscribed text on themonitor.

The external USBsound card shouldhave a blinking blueLED. Any detectedwords should outputto the terminal. Thekeywords will bedetected in the orderthey are spoken.

Once the programis launched, theLED on the USBadapter begins toblink. In theterminal, alistening prompt isshown, along witha spinning bar.Any spoken wordsare showntranscribed in theterminal output.

Pass

4 Run the motor controlprogram. Verifyfunctionality offorward and reversevoice commands. Runthe program for 5minutes, varyingdirection every 30seconds.

Both motors shouldrespond accordinglyto voice commands.When a reversecommand is given,both motors shouldspin in the oppositedirection.

When given aforwardcommand, bothmotors begin tospin forward andthe chair begins tomove.

Pass

5 Run the motor controlprogram. Verifyfunctionality of turningcommands. Run theprogram for 5minutes, issuemultiple left and rightturn commands.

Both motors shouldrespond to the voicecommands. When aleft turn is issued, theleft wheel will go inreverse, while theright wheel goesforward.

Once a commandis issued, bothmotors spin in thecorrect direction.This causes ourchain to spin andpropel our chairforward.

Pass

6 Run the motor controlprogram. Verify thefunctionality of thestop command. Whilemoving forward, issuea stop command. Doanother test while inreverse.

Once the stopcommand is issued,the motors shouldcome to a stop in lessthan 3 seconds.

The motorsrespond to stopcommands whenissued. Anotheralternativecommand is ‘kill’.

Pass

7 Run the motor controlprogram. Verify thefunctionality of thestop command. Whileperforming a turn,issue a stopcommand. Do a testwhile turning left andright.

Once the stopcommand is issued,the motors shouldcome to a stop in lessthan 3 seconds.

Once the stopcommand isissued, both of themotor’s come to acomplete stop in3-5 secondsdepending on thespeed.

Pass

8 Run the motor controlprogram. Verify thefunctionality of everycommand. Startmoving forward, domultiple turns andreverse movements.Verify the stopcommand is functionalwhile performing anymovement.

The wheelchairshould respond toany definedcommand given bythe test user. A stopcommand shouldbring the chair to ahalt at any time.

Once a commandis issued, themotors respondcorrectly. Thechair has theability to moveforward, reverse,turn left or right.This is behavingas intended.

Pass

9 The Motor Driver(VESC) powers onwhen 12 Volts isapplied to it.

The LED indicatingpower-on emits agreen light

Once power isapplied to aVESC, the greenLED isilluminated.

Pass

10 The Motor Driversends a signal viapulse-widthmodulation to themotors

The motor shaftrotates at the rpmspecified by ourmotor control.

Once a RPMcommand is sentto a VESC, themotor spinsaccordingly.

11 The 12 V DC Motorshaft rotates when 12volts is applied tothem

The motor shaftrotates at max rpm

This test is notapplicable, as weare using BLDCmotors insteadnow.

N/A

12 The wheels areproperly attached tothe motor shaft

The wheels rotate atthe same rpm as themotor shaft

After grinding flatspots for our setscrews on wheelsand axle sprocket,we confirmed ourtorque istransferring asexpected.

Pass

13 Both VESCcommunicate withRaspberry Pi 4 tool

When a VESCreceives a signalfrom the RPI, thecorresponding motoris powered.

Once a commandis sent to theVESC, the motorspins up to the setRPM.

Pass

14 The Motor Driver iselectrically connectedto the motors

A voltage can beseen at all the pointsof contact betweenthe Motor Driver andleads to the motors

The motors areconnectedcorrectly, andreceive the correctsignals.

Pass

15 The Motor shaftrotates bi-directionally

The motor shaftrotates clockwisewhen a positivevoltage is applied andcounterclockwisewhen a negativevoltage is applied

After calibratingthe motors withthe VESC tool, weset the motororientation. Themotors spin in thecorrect direction.

Pass

16 Check distance forlidar by inputting 150

cm into code and thenmeasuring the

distance when themotors stop running.

Our expected resultis 4.9 feet. Motors

should stop when anobject is detected

less than or equal to4.9 feet

Motors stop whenan object is lessthan or equal to

4.9 feet. Tested byusing the hand

and thenmeasuring the

Pass

distance betweenthe sensor and

hand.

17 Check voltage for lidarusing voltmeter

Voltage for lidarshould be 3.3V

Voltage for lidar is3.3V

Pass

18 Check stop time afterdetection

Stop time should beinstant

Once a speed ofzero is applied,

the motors brakeinstantly.

Pass

19 Check if lidar changesthe key value by

displaying key valueeach time it changes

Lidar should changethe value to 8

Lidar changesvalue to 8

Pass

20 Check if lidar changesthe state value by

displaying state valueeach time it changes

Lidar should changethe state value to 0

Lidar changesstate value to 0

Pass

21 Check if lidar isrunning with other

devices

Lidar should displayan object detect text

string.

When object is infront of sensor, it

displays text string

Pass

22 Check if lidar isalways running by

inputting a text stringunder the code to seeif it is being accessed.

Text string shoulddisplay “Lidar

Running”

Lidar sensorcontinuously

running, it showsthat the Lidar isRunning via text

string

Pass

23 Check if lidar haspriority over voicecommands and

motors

Lidar should be ableto stop the motorsinstantly when an

object gets detected

Lidar has priorityand can stop

motors instantly

Pass

24 Check If motor onereceives forward state

1 by issuing thecorrect voice

command in pi

Motor 1 is goingforward

Motor goesforward

Pass

25 Check If motor tworeceives forward state

1 by issuing the

Motor 2 is goingforward

Motor goesforward

Pass

correct voicecommand in pi

26 Check If motor onereceives reverse state

2 by issuing thecorrect voice

command in pi

Motor 1 is going inreverse

Motor goesreverse

Pass

27 Check If motor tworeceives reverse state

2 by issuing thecorrect voice

command in pi

Motor 2 is goingreverse

Motor goesreverse

Pass

28 Check if both motorsstop by issuing the

correct vice commandin pi

Motor 1 & 2 both stopinstantly

Motor 1 & 2 bothstop instantly

Pass

29 Check if motors moveforward while on the

ground

Both motors moveforward at same

speeds, resulting inthe chair moving

forward

The chair is ableto move forward ina straight fashionwhile bearing low

weight.

Pass

30 Check if motors movein reverse while on

the ground

Both motors move inreverse at same

speeds, resulting inthe chair moving

forward

The chair is ableto move in a

reverse fashion.

Pass

31 Check state valuelogic by printing textstrings for each statein order to check andsee if each state is

being accessed at thecorrect time

Text stringcorrespond correctly

to the state value

Each state isbehaving as

intended. When acommand isissued, the

appropriate actionis taken.

Pass

32 Check if lidar, voice,and motors are

running at the sametime by printing textstrings within each

The text stringsshould correspond toeither voice, lidar, or

the motors

Each process isrunning as

expected. When acommand or

object is noticed,

Pass

multiprocessor action is taken.

33 Check if universal keyvalue is correctlybeing updated bydisplaying value

The key value shouldcorrespond with the

voice commands

Once a definedcommand is

issued, the keyvalue is updated.

Pass

34 Check u value bydisplaying the value

and seeing if itmatches the

MAX_SPEED

U value should be setto MAX_SPEED

Once max speedis reached, our uvalue is set. Thechair will remain

at max speed untila new commandis issued or an

object isencountered.

Pass

35 Check if multipleprocesses are runningat same time. Check

the running processeswith the topcommand.

It should display thedifferent processes

runningsimultaneously

The differentprocesses areseen running

alongside eachother.

Pass

IX. Marketability Forecast

In our initial research into thisproblem, we discovered that spinal cordinjuries have been estimated to beprevalent in 200,000 to 288,000 peopleliving in the United States. The higherup the spine the injury takes place, moreof the body is affected. Withquadraplegic and paraplegic patients, theability to move their head and facemuscles remains.

Alternate approaches to thisproblem for wheelchair control haveused head motion to identify usercommands. This type of control wouldnot work if they are unable to move theirneck. A researcher at North CarolinaState University has designed a TongueDrive System(TDS). This allows theuser to control their wheelchair bymoving their tongue. This is done byputting a metal piercing on the person'stongue, and fixing a small sensor insideof the mount. This technology isdescribed by Patent US8044766B2. Thisapproach requires implanting hardwareinside the user's mouth, which isn’talways possible.

Figure 14 - Tongue drive system

Another design approach hasbeen described in Patent US5812978A.The commands described andcapabilities are similar, with a few keydifferences. Our design incorporates aLidar sensor to detect objects in our pathand stop when necessary. Their designdoes not include such a feature. Thepatent also describes using a throatmicrophone (laryngophone) compared toour standard sound wave microphone.This would allow the system to detectcommands by vibrations of the throat,effectively ignoring any backgroundnoise. The downside of this approach isthis technology is not widespread andcompatibility with Unix may vary.

Figure 15 - Image showing a person pushing awheelchair by themself. [19]

In our research, our wheelchair istaking the regular wheelchair andexpanding the ability for mobility. Theproblem with a regular wheelchair is thatin order to get mobility you must havesomeone who is willing to push thewheelchair for you, or be able to do ityourself which requires you to be able touse your arms and hands. This works forpeople that are unable to move but havefunction for their hands and arms butthis does not work for someone that haslost the ability to use their hands orarms. As shown in figure 13, in order fora person to push themself in a regularwheelchair they must have function oftheir arms and hands.

Figure 16 - A person pushing disabled person[20]

Another issue is that in order tohave mobility another person must bepresent in order to give you mobility onthe wheelchair. As shown in figure 15,another person is needed in order to pushthe person in the wheelchair. This limitsmobility because the disabled person'smobility depends on when anotherperson is available.

Figure 17 - A person pushing disabled person[21]

The electric wheelchair isanother competitor in the market. Theyare an option to a disabled person thatdoes not have function to their arms buthas function to their hands. The issuewith electric wheelchairs is that it stillrequires use of hands in order to movethe wheelchair by using the joystick. Ourwheelchair allows the person to controlthe wheelchair with just their voicewithout needing another person the useof hands or arms.

SWOT (Strengths, Weaknesses,Opportunities, Threats) Analysis:

Strengths of the product includeits usability, efficiency, andcomfortability. Our wheelchair benefitsfrom being voice-controlled, whichgives it a lower barrier-to-entry for userswith a wide range of disabilities. Ourwheelchair benefits from being electricpowered, which makes it more efficientthan a manually operated wheelchair. Auser of our wheelchair could cover twicethe distance in half the amount of time asa user of a non-electric product. Ourwheelchair benefits from beingergonomic. The use of a microphone thatis positioned near the mouth to speakinto, as opposed to a larynx microphonethat is positioned at the throat to pick upvibrations, is more comfortable forusers. Additional strengths of the projectinclude the use of low-cost materials andtechnologies for easy assembly, such aspre-build microcontrollers andmotor-drivers, along with an aluminum

frame. On top of all that, our teampossesses skilled computer and electricalengineers that have developed the codeso the chair adapts to its environment.The chair stops when obstacles arepresent, and in the future our engineerscould develop a method to make thewheelchair self-driving.Weaknesses of the product include safetyrisks and potentially high-cost ofmanufacturing. Our wheelchair currentlylacks a seatbelt for securing the user inthe seat. This is dangerous, especially ifthe user is accelerating at a speed ofthree to five miles per hour. Ourwheelchair also could be costly toacquire. The average consumer wouldlikely not be able to afford the initialversion of the product without a way tosubsidize the cost.Opportunities of the product includepatentability, licensing, and adaptabilityof the control system to various models.Our wheelchair is uniquely designed.This opens the door for patentabilitywith the United States Patent andTrademark Office (USPTO). TheUSPTO allows the registration of designpatents that show novelty in theorientation or arrangement of aninvention. Our wheelchair’s system isengineered using an assortment of manytechnologies that make it a novelrepresentation of a wheelchair. With apatent, we could then license our designto wheelchair manufacturers. The cruxof our product would be its controlsystem. There are many existingproducts that would be improved by

equipping a voice-control drive system,i.e. cars, shopping carts, bikes, androbots.Threats to our product include marketoversaturation and obsolescence. Ourwheelchair is essentially a rendition ofan already existing product. There aremany models of wheelchairs on themarket, at different price points, whichmakes the possibility that our productwould be lost in all the noise very high.To mitigate this, our team would work tomake our product stand out by includingadditional features in an updated model,such as bluetooth connectability,smart-device integration, and batterycharging. Making our product moretechnologically compatible, would allowour customers to integrate the currentdevices they use on a regular basis withour wheelchair, making it moreattractive. The additional upgradeswould likely keep our product in themarket long enough before the secondthreat completely makes it obsolete.New technologies are being developedevery day, so the use of a wheelchairmay very well become a thing of the pastas medical science advances. Once thathappens, there is really no way to alterour product to address the needs of ourcustomers.

X. ConclusionA. Societal ProblemPeople with disabilities exist.

They are a large part of our high-techsociety and therefore they deservehigh-tech solutions that make their liveseasier. The wheelchair is one of thosesolutions, yet its current incarnation doesnot address the needs of all people withdisabilities. There are many types ofwheelchairs, differing by the method ofpropulsion, control mechanism andtechnology used. Some wheelchairs aredesigned for general everyday use,others for single activities or to addressspecific accessibility needs. Innovationin the wheelchair industry is common,but many innovations end up in a state ofunderdevelopment, either byover-specialization, or by not bringingan idea to market at accessible prices.Our project is an introduction to the nextgeneration of wheelchairs.

B. Design IdeaOur design prototype is going to

implement various sensors, with the goalof making a voice-controlled wheelchair.Safety is going to be a primary concern,so certain features will need to beimplemented. We plan to include amicrophone in our system, which willallow our users to input voicecommands. These commands will allowthe wheelchair to be controlled handsfree. This will allow disabled people togo about their day-to-day life. We planto focus on core functionality for ourinitial prototype, with the plans to addadditional control features in the future.

The Raspberry Pi onboard will allow usto implement other hardwarecomponents if necessary. Open-sourcelibraries will also allow us to easilycommunicate with any other sensors wemay need.

We are envisioning our design tobe implemented in a hospital typeenvironment, where patients are oftentransported with personnel assistance.This technology would allow hospitalworkers to focus on other tasks at hand,which could improve efficiency in dailyoperations. With employee shortagesbeing seen in most sectors, thistechnology makes sense from a coststandpoint. Technology allows us totackle challenges that would otherwisebe impossible. If this device can makesomeone’s life a little easier, it willprovide a great deal of value.

C. Work Breakdown StructureThis project consists of various

assignments that need to be completedby their respective due dates to meet thedelivery date of May 2022. The WorkBreakdown Structure Table, shown inAppendix D, is a detailed list of allassignments needed to be completed tomeet the requirements that make up eachfeature of the prototype. Each teammember has accepted assignments thatrelate to the designing and building ofthe prototype features based on theirskills and capabilities. This guaranteesthat each feature will be completed withefficiency and competency, on time fordelivery.

D. Project TimelineThis timeline will be a tool that we

utilize and maintain through the designprocess. Visualizing our assigned taskswill allow us to align our workload withthe assignments, ensuring we fulfill therequirements. Defining our workpackages and assigning a timeline willensure any bottlenecks or complicationsin our design will be managed in atimely fashion. Major Milestones aredefined as our prototype presentations.

E. Risk AssessmentThe risk assessment matrix is a tool

we used to identify any risks that we as ateam might face. The risks taken inaccount were environmental risks andproject risks. Having a risk matrixallowed us to develop mitigations foreach risk in our risk matrix. This allowsus as a team to make sure that if we dorun into any problems, we have a backupplan to solve that problem. For each riskon our matrix, we have the probability ofthe risk occurring.

G. Device Test PlanThe test plan goes through each elementin our feature set, and demonstrates thefunctionality of our system. Testing eachindividual component initially, thisallows us to confirm the operation of ourdevices. Once we confirm the pieceswork, we run tests which incorporate theentire system. These tests include issuingvoice commands and verifying themovement of our motors. At any time,

the stop or ‘kill’ command which brakesboth motors.

H. Market ReviewThis project is for the purpose of

testing and exploring. Thereforefocussing on market analysis at thistime is not an urgent step. In the futurewhen we have completed the finalproduct review, another market reviewwill be executed before the product islaunched to the customer's market. Afterthe review, we have decided the targetedcustomers, price analysis and level ofcustomers’ needs.

I. Testing ResultsAfter the testing, team 13 has

completed the initial objectives whichwas listed in the punch list as: Controlwheelchair go forward, reverse, turn left,turn right and adjust the wheelchairspeed through adjusting RPM andvoltage supplied. Compared to ourprevious model of PVC wheelchair,there are some improvements which wewere able to achieve in this semester.With the new frame we were able tosecure the wheelchair with a heavy loadon it. The new motors were strongenough to move the wheelchair withheavy load, this solved our torque issuewhich we mentioned at the beginning ofsecond semester. The Lidar sensor wasable to recognize objects in front of thewheelchair and stop the wheelchairbefore collision.

In addition to the obtained results,the model has the following limitations:

There is still a delay in the controlfrom voice control to the motor. Voicecontrol is not optimized, the delayshappen while receiving and transcribing.data. This can be because of theobjective factors. Our speech to text isrunning a tensorflow lite model, which isa CPU intensive task. With futureimprovements in speech models andcomputing power, this could be a morereliable method of control.

The lidar sensor is working well sofar when there's an object detected infront of the wheelchair. However tooptimize the ability of collisionavoidance, the wheelchair needs to beable to detect objects from everydirection. This could be improved byadding more lidar sensors at differentpositions of the wheelchair.

The wheel is not currently receivingall of the torque produced by the motor.During testing, our drive system wasable to move up to 35 lbs of additionalload. Once we attempted to add another10 lbs, the sprocket began to slip on themotor shaft. This is due to the fact weare using a sprocket that is 2mm toolarge for our motor shaft. Due to thisshipping error and delayed replacement,we attempted to fix the issue byincreasing our motor shaft diameter withaluminum foil tape. This issue should beresolved with a proper fitting sprocket.

REFERENCES

Article Journal Citation

] 1 Human-machinecommunication

Computer in HumanBehavior

Volume 90, January 2019,Page 285-287

Patric R. SpenceHuman-machine communication

https://www.sciencedirect.com/science/article/pii/S0747563218304540

2 BASICS OF UARTCOMMUNICATION

Circuit Basis Scott Campbell,Basics of Uart communication

https://www.circuitbasics.com/basics-uart-communication/

3 Wheelchair History Wheelchair History https://en.wikipedia.org/wiki/Wheelchair

4 History of theWheelchair

History.Phiso Glenn Ruscoe,History of wheelchair

https://history.physio/history-of-the-wheelchair/

5 Who was the firstindividual to invent the

wheelchair?

Sport News Sports NewsJanuary 29, 2021

https://sportsnewsstar.blogspot.com/2021/01/who-was-first-individual-to-invent.html

6 Narrow Wheelchairs forTight Spaces and Skinny

Doorways

Grayingwithgrace.com Scott Grant

https://www.grayingwithgrace.com/wheelchairs-narrow-doorways/

7 Jazzy 600 ES Spinlife https://www.spinlife.com/Pride-Jazzy-600-ES-Ful-Size-Power-Wheelchairs/spec.cfm?productID=103386&adv=googlepla&utm_medium=CSE&ut

m_source=googlepla&utm_term=&utm_campaign=11097663967&gclid=Cj0KCQjwnoqLBhD4ARIsAL5JedL82bvVIob0x2ehL23l8kgU8tqBwk7Ld

DPNllz9QrUT__MlcEEfOYAaArx-EALw_wcB

8 Draco (Electric StandingWheelchair)

Living Spinal https://livingspinal.com/active-mobility/standing-wheelchairs-and-frames/draco-electric-standing-

wheelchair/

9 Designing the world’sfastest Wheelchair

Viterbi Magazine Daniel Druhora

https://magazine.viterbi.usc.edu/spring-2017/alumni/designing-the-worlds-fastest-wheelchair/

10 Arduino basedvoice-controlled

wheelchair

IOP Science Tan Kian Hou et al 2020 J. Phys.: Conf. Ser.1432 012064

https://iopscience.iop.org/article/10.1088/1742-6596/1432/1/012064/pdf

11 Voice ControlledIntelligent Wheelchair

usingRaspberry Pi

International journal ofengineering research and

technology

Naeem, A., Qadir, A., & Safdar, W. (2014).Voice Controlled Intelligent Wheelchair

using Raspberry Pi. International journal ofengineering research and technology, 2.

12 Control Systems andElectronic

InstrumentationApplied to Autonomy

in WheelchairMobility:

The State of the Art

Sensors 2020, 20(21), 6326 Callejas-Cuervo M, González-Cely AX,Bastos-Filho T. Control Systems andElectronic Instrumentation Applied to

Autonomy in Wheelchair Mobility: The Stateof the Art. Sensors. 2020; 20(21):6326.

https://doi.org/10.3390/s20216326

13 Low-cost sensornetwork for obstacle

avoidance inshare-controlled smart

wheelchairs underdaily scenarios

Journal article Pu, Jiangbo & Jiang, Youcong & Xie, Xiaobo& Chen, Xiaogang & Liu, Ming & Xu,

Shengpu. (2018). Low cost sensor networkfor obstacle avoidance in share-controlledsmart wheelchairs under daily scenarios.

Microelectronics Reliability. 83. 180-186.10.1016/j.microrel.2018.03.003.

14 How many peoplewould benefit from a

smart wheelchair?

Journal Article Richard C. Simpson, PhD, ATP; Edmund F.LoPresti, PhD; Rory A. Cooper, PhD,

Department of Rehabilitation Sciences andTechnology; University of Pittsburgh, PA,

JRRD Vol. 45 pgs. 53-72

15 Disability Impacts Government website https://www.cdc.gov/ncbddd/disabilityandhealth/infographic-disability-impacts-all.html

116 Hacking the brain:brain-computer

interfacing technologyand the ethics ofneurosecurity.

Ethics and InformationTechnology

Ienca, M., Haselager, P. Hacking the brain:brain–computer interfacing technology and

the ethics of neurosecurity. Ethics InfTechnol 18, 117–129 (2016).

https://doi.org/10.1007/s10676-016-9398-9

17Hate crime or mate

crime? Disablisthostility, contempt,and ridicule: Pam

Thomas

Book Thomas, P. (2012). Hate crime or matecrime? Disablist hostility, contempt, and

ridicule: Pam Thomas. In Disability, HateCrime and Violence (pp. 140-151).

Routledge.

18 Head-motioncontrolled wheelchair

Journal Article S. Prasad, D. Sakpal, P. Rakhe and S.Rawool, "Head-motion controlled

wheelchair," 2017 2nd IEEE InternationalConference on Recent Trends in Electronics,Information & Communication Technology

(RTEICT), 2017, pp. 1636-1640, doi:10.1109/RTEICT.2017.8256876.

19 How to use aWheelchair

Webpage https://www.wikihow.com/Use-a-Wheelchair

20 Wheelchair Patient Webpage https://www.clipartmax.com/middle/m2i8i8N4b1b1d3Z5_wheelchair-patient-nursing-cli

p-art-push-wheelchair/

21 Electric Wheelchair Webpage https://www.kindpng.com/imgv/iTJioxm_null-electric-wheelchair-clip-art-hd-png-downl

oad/

Glossary

BCM: Broadcom chip-specific pin numbersPWM: Pulse width modulated

Coqui STT: Coqui Speech-to–text

NEED TO FILL IN

Appendix A. Hardware

NEED TO FILL IN

Appendix B. Software

NEED TO FILL INAppendix C. Mechanical Aspects

Appendix D. Work Breakdown Structure

Level 1 Level 2 Level 3

Voice Control

Wheelchair must becapable of navigating inresponse to voicecommands.This task will be assignedto Erik Coleman, with helpfrom the group for testing.Shazib will assist if furtherimplementation is neededfor speech recognition.

Subtask 1.1:Stream user audio to voicerecognition software.

Subtask 1.2:Take user commandsdetected from speechrecognition software tosend to our motorcontroller.

Work Package 1.1.1:Setup microphone andverify correct functionality.Time estimated: 2 hoursStatus: Complete

Work Package 1.1.2:Verify speech recognitionsoftware provided with themicrophone will work forour purpose. Onceconfirmed, included in thefinal design.Time estimated: 4-6 hoursStatus: Complete

Work Package 1.2.1:Develop python code totrigger an action in ourmotor controller, based onuser input from themicrophone.Time estimated: 4 hoursStatus: Complete

Work Package 1.2.2:Test our voice controller byusing this functionality in areal-world environment,with multiple users. Whenthis feature is working,demonstrate ability in theprototype presentation.Time estimated: 2-4 hoursStatus: In progress

Electric Drivetrain

Wheelchair will bepowered by an electricbattery. The 12V batterywill provide power to ourDC motors, allowing thewheelchair to maneuveritself.This task will be assignedto Waverly, with assistancefrom Pho with circuitconstruction. The entireteam will be involved withtesting.

Subtask 2.1Power motors using anelectric battery

Subtask 2.2Build frame to house thebattery

Subtask 2.3Use motors to rotate twoback wheels

In the new version of theproject for second semesterwe replaced the old motors12V with the built-in gearbox motor 24V to increasethe torque power.Encoder will be an add-onapplication which willassist Shazib and Erik withthe programming.

Subtask 2.4New motor with built-ingear box will increasetorque power

Subtask 2.5Encoder application isexplained in Work Package6.1.3

2.4.1Swapping out the oldmotors with new motorsStatus: In Progress

Battery will be upgraded to24V. Our plan is to connecta new 12V battery in serieswith the old battery whichwe have right now in casewe need.

Subtask 2.6The series of 12V batterywill supply stronger powerto run the two new motors

2.6.1We will test the newversion of the wheelchairwith a 12V battery. In caseit doesn’t supply enoughpower, the series of 12Vbatteries will be a back-upplan.Status: Testing

Work Package 2.1.1Research parts andsuppliers for motors,batteries, and wheels.Time estimated:Status: In progress

Work Package 2.1.2List parts in Assignment 3– Design ApproachTime estimated:Status: In progress

Turning

Wheelchair must have aturning radius of 360degrees.

Subtask 3.1Design a method for thewheelchair to turn.

Work Package 3.1.1Calculate rotation speedneeded to turn wheels.This task will be assignedto Waverly and Pho.Time estimated: 2-4 hoursStatus:In progress

Work Package 3.1.2Program motor-driver toregulate motor speed.This task will be assignedto Erik.Time estimated: 2-4 hoursStatus: Complete

Collision Avoidance

The wheelchair must becapable of sensing objectsaround itself. Once an

Subtask 4.1 A lidar sensorwill be used to detectobjects and humans

object is determined to bein our path, we must stopand wait for user action.

Work Package 4.1.1A lidar sensor is to beattached to the wheelchairand connected to the Rpi4.Set up the lidar sensor todetect objects within a 5ftradius. Wheelchair stopswhen an object is detected.This task will be assignedto Shazib.Time estimated: 2-4 hoursStatus: Complete

Work Package 4.1.2Verify the sensor operatescorrectly in a crowdedenvironment with multipleobjects at a time. This taskwill be assigned to theentire team, withcontinuous testing.Time estimated: 4 hoursStatus: In progress

Forward and ReverseAcceleration

The wheelchair must becapable of accelerating in aforward and reversemotion.

Subtask 5.1 Test motordriver functionality withour electronics and motors.

Subtask 5.2 Test motorfunctionality whileattached to our prototypeframe and wheels.

Work Package 5.1.1:Attach the motor driver toRaspberry Pi, power themotor driver with the 12Vbattery. Power the RPI witha separate battery pack.This task will be assignedto Waverly and Pho.Time estimated: 2-4 hoursStatus: Complete

Work Package 5.1.2:Test the motor driverfunctionality by sending aPWM signal, in bothforward and reversedirection. Verify the motorsswitch direction.This task will be done bythe entire team. Erik andShazib will focus on thesoftware.Time estimated: 2 hoursStatus: Complete

Work Package 5.2.1:Assemble the frame for theprototype, securing ourelectronics to our chosenframe. This task will becompleted by the entireteam.Time estimated:4-6 hoursStatus: Complete

Work Package 5.2.2:Test our prototype bysending forward andreverse commands,verifying the correct

movement is taken. Oncethis is verified, our designis ready for TechnicalEvaluation (Assignment 7).This task will be completedby the entire team.Time estimated: 2 hoursStatus: Complete

Speed Detection

Wheelchair must becapable of detecting andmeasuring its own velocity.

Subtask 6.1An encoder is used tomeasure the movingvelocity of a motor shaft.

Work Package 6.1.1Find a supplier for a motorwith encoder, and thenpurchase it. This task willbe assigned to Pho.Time estimated: 2 hoursStatus: Complete

Subtask 6.2Pulse Width Modulationadjusts the output voltageof the load based on thechange in the width of thepulse sequence. This willallow us to adjust speed asneeded.

Work Package 6.1.2Determine correct methodto adjust PWM duty cycleto allow the user to varythe wheelchair speed. Thistask will be done by theentire team.Time estimated: 2 hoursStatus:In progress

Work Package 6.1.3Implement a PID controlloop to adjust motor speed.Our goal is to minimizedrift and to maintain aconsistent speed. The erroris calculated between ourtarget speed and actualspeed, which is then usedto adjust the motors voltageaccordingly. An encoder isused to provide samplevalues from our motor.These ticks are used tocalculate error. This will beimplemented by Shaziband Erik.Time estimated: 4 hoursStatus:In progress

Figure 2: Flowchart diagram illustrating work breakdown structure.

Appendix E. Timeline Charts and PERT Diagrams

PERT Diagram

Appendix F. Resumes

Erik Coleman

Employment HistoryCalifornia Department of Transportation (June 2018 – August 2019)Student AssistantDatabase specialist· Implemented several databases to record incidents and analyze data.· Created training presentations documenting safety procedures and protocols.· Various office tasks (Inventory, scanning, filing, etc.)Northern California Power Agency (N.C.P.A) (August 2019 – October 16. 2020)IT Systems Administration Intern· Setup machine and workstation following company standards.· Manage Active Directory and Azure Active Directory users and groups for internaland extranet permissions.· Write PowerShell scripts to automate tasks and track machine health.· Assist users remotely with troubleshooting network and device issues.

SkillsProblem SolvingTroubleshootingRTL design and verificationDigital circuit analysisAccess database designNetwork configurationDocker container managementLanguages:CVerilogV.H.D.LPowerShellC++PythonAssembly

Projects· At Sacramento State, I was a part of a team responsible for designing and simulatinga 5-stage pipelined processor in Verilog. We programmed it modularly and ran testbenches verifying each component in simulation.

· While working at N.C.P.A, I configured a deployment server to automate Windows10 O.S. deployment and software installation using MDT/WDS over PXE boot.· I programmed a traffic light state machine in VHDL while at Yuba College.· While at Yuba College, I programmed a 5-axis robot to draw on a whiteboard. Thecode was written in C++ and converted into G-code. It was then outputted over serial tothe mechanical robot.

Waverly Hampton III, EIT

Employment HistoryAmbulnz Health, LLC + First Rescue Ambulance (December 2016 – October 2017)Emergency Medical Technician

● Assessed, treated, and transported the sick and injured to required medical facilitiesPrivate Tutor + Chaffey College (October 2017- July 2019)Math & Physics Tutor

● Tutored grade school, high school, and college students in mathematic subjects,ranging from Algebra to Differential Equations. Included Physics lessons for certainclients, which involved building model rockets and battery-powered toy cars.

Start-up (August 2019-Present)Chief Executive Officer

● Manage operations and execute strategies for small businesses

SkillsOffice: MS Word, Powerpoint, Excel, Google Docs, Pages, Numbers, Keynote Virtual: Zoom,MS Teams, Google Meet, FaceTimeDesign: Solidworks, Adobe Photoshop, Adobe Illustrator, Adobe Dreamweaver, Adobe XD,AdobeLanguages: Python, C++, Java, HTML5, Swift, Pascal

Shazib Naveed

Employment History Cell Phone Shop (December 2017 – October 2019) Repair Technician & management

● Repaired cell phones and figured out solutions to software related problemsfor customers. ● Managed inventory of the cell phone shop

Small business (January 2020- June 2020) Online computer & cell phone repairs

● Maintained an online shop using various social media to help customers withphones and computers online.

Volt (August 2021-Present) Cell Phone Repair Engineer

● Working with apple phones to fix and repair hardware or software relatedissues.

Skills Office: MS Word, PowerPoint, Excel, Google Docs Languages: C, C++ Verilog V.H.D.L Python Assembly Java HTML

Pho V Le

Employment History Assistant Store Manager, O’reilly Autopart, Florin Rd, Sacramento CASept 2020-Sept 2021

● Assisting the district manager in supervising and scheduling store team members. ● Responsible for understanding cost control, the store's P&L, and how to operatea profitable store ● Supervise the professional and retail operations of the store and team membersinvolved ● Coaching new team members ● Learn to manage key components of gross profit ● Monitoring/reinforcement of safety expectations

Sales Manager, AutoZone Inc., Fruitridge Rd, Sacramento CA May 2018 –Aug 2020

● Assisting the store manager in supervising and scheduling store AutoZoners. ● Assisting customers in finding parts and products using electronic or paper catalogs. ● Operating cash registers and following established cash handling duties, includingbut not limited to deposits, petty cash, and lane accountability. ● Ensuring store merchandising tasks, including but not limited to stockingthe store, are completed in a timely and accurate manner. ● Processing returns and managing inventory. ● Conducting and reviewing all opening and closing procedures.

Customer Services, AutoZone Inc., Fruitridge Rd, Sacramento CA Sep 2017 –May 2018

● Operated cash registers and followed established cash handling procedures ● Ensured that merchandise was restocked and placed in their respective areas. ● Helped customers to diagnose automobile problems and recommended solutions.

● Provided customers with product knowledge and current promotions throughAutoZone systems. ● Communicated with managers regarding customer concerns and employee matters.

Skills ● MS Word, PowerPoint, Excel, Google Docs ● Programming language (C++, Python). ● Photoshop using Adobe Acrobat

Projects ● Building Traffic light using Microcontroller (STM32 Nucleo F303K8) ● Learning UART/USART and applying in building Project (throughMobaXterm) ● Build optocoupler circuit, apply sensor in the circuit and simulation. ● Integrate system with a Heartbeat Measurement Circuit. ● Building “wheel of fortune” arcade game. An arcade with 3 lights and a button sowhen the light spins in a circle which every light was landing on after hitting the bottom willtell us how many tickets we will get. (Raspberry Pi, STM32 and Finite State Machine wasapplied)