BRAIN CONTROLLED ARTIFICIAL LEGS

FROM:

Engr Rana Muhammad Shakeel

Engnr.shakeel@gmail.com

https://www.facebook.com/EngnrShakeel

plz like my page:

https://www.facebook.com/Electrical4Electronics

For MORE PROJECTS:http://electro-technolgy.blogspot.com

1

ABSTRACT: This paperdescribes a brain controlled robotic leg whichis designed to perform the normal operations ofa human leg. After implanting this leg in ahuman, the leg can be controlled with the helpof user’s brain signals alone. This leg behavessimilar to a normal human leg and it can performoperation like walking, running, climbing stairsetc. The entire system is controlled with thehelp of advanced microcontrollers and digitalsignal processors. The signals are taken outfrom the human brain with the help ofelectroencephalography technique. The person can

2

perform operations like walking, running etcjust by their thought. This system will be verymuch suitable for those who lost their legs inaccidents and the proposed system is hundredpercent feasible in the real time environmentwith the currently available technology.  The BrainControlled Artificial Legs are very much costeffective when compared to the normal Artificiallegs which is available in the market. Thereduction in cost of the proposed system isfound to be above 80% when compared to theexisting system. Moreover, the user canhave full control over the artificial legs whichis not possible in the existing system.

Introduction:

Abrain-computer interface (BCI), sometimes calleda direct neural interface or a brainmachineinterface, is a direct communication pathwaybetween a human or animal brain and anexternaldevice. In this definition, the word brain meansthe brain or nervous system of an organiclifeform rather than the mind. Computer means anyprocessing or computational device, fromsimplecircuits to the complex microprocessors andmicrocontrollers.An interesting question for the

3

development of a BCI is how to handle twolearning sys -tems: The machine should learn todiscriminate between different patterns of brainactivity as accurateas possible and the user ofthe BCI should learn to perform different mentaltasks in order to produce distinct brainsignals. BCI research makes high demands on thesystem and software used. Parameter extraction,pattern recognition and classification are themain tasks to be performed in a brain signals.In this paper it is assumed that the user ofthis system has one leg which is functioningfully and the system is designed accordingly.This system can be extended for both the legsand it is not limited to the basic operation ofhuman legs such as walking, running, climb - ingstairs etc. It can also perform operations likecycling, hopping etc.

Brain Waves:

Electrical activity emanating from the brain isdisplayed in the form of brainwaves. There arefour categories of these brainwaves ranging fromthe most activity to the least activity. Whenthe brain is aroused and actively engaged inmental activities, it generates beta waves.These beta waves are of relatively low

4

amplitude, and are the fastest of the fourdifferent brainwaves. The frequency of betawaves ranges from 15 to 40 cycles a second. Thenext brainwave category in order of frequency isAlpha. Where beta represented arousal, alpharepresents non-arousal. Alpha brainwaves areslower and higher in amplitude. Their frequency ranges from 9 to 14 cycles persecond. The next state, theta brainwaves, istypically of even greater amplitude and slowerfrequency. This frequency range is normallybetween 5and 8 cycles a second. A person who hastaken time off from a task and begins todaydream is often in a theta brainwave state.The final brainwave state is delta. Here thebrainwaves are of the greatest amplitude andslowest frequency. They typically center arounda range of 1.5 to 4 cycles per second. Theynever go down to zero because that would meanthat you were brain dead. But, deep dreamlesssleep would take you down to the lowestfrequency. Typically, 2 to 3 cycles a second.

5

Different Types of Brain Waves

In the proposed system alpha waves and betawaves are used from the brain for signalprocessing.It is assumed that the person is inalpha state and beta state (which is the casenormally) and these waves are taken out from thehuman brain and converted in the form ofelectrical signals with the help of electrodecaps. The following figure shows the differenttypes of waves and also the mental state of theperson.Those waves usually vary from afrequency of 1Hz to 40 HZ.

General Block Diagram of the System:

6

General Block Diagram of the Proposed System

Fig shows the generalblock diagram of the proposed system. Electrodecap isplaced in the scalp of the person. Thesignals taken out from the human brain will bein the range of mV and μV. Hence they are fed toan amplifier. Then it is sent to a Analog toDigitalConverter to convert the analog brain signals into digital form. Then it is sent to a signalprocessor where parameter extraction, patternclassification and pattern identification aredone. These digital signals are fed as input tomicrocontroller unit. The last four units(Amplifier, Signal Processor, Analog to DigitalConverter and Microcontroller Unit) are placedinside the artificial leg. The output of themicrocontroller unit is fed to the drivingcircuit. Let us see about these blocks indetail.

7

Electrode Cap:

Fig3 shows a person wearing an electrode cap.These electrode caps contains electrodes whichare placed on the skull in an arrangement called10-20 system, a placement scheme devised by theinternational federation of societies of EEG. Inmost applications 19 electrodes are placed inthe scalp. Additional electrodes can be added tothe standard set-up when a clinical or researchapplication demands increased spatial resolutionfor a particular area of the brain. High-densityarrays (typically via cap or net) can contain upto 256 electrodes more-or-less evenly spacedaround the scalp. The main function of theelectrode cap is to take the brain signals inthe form of electrical signals. The signalstaken out from the Electrode cap are fed to anamplifier.

8

Fig:A person wearing electrode cap fig: Placement of electrodes in 10-20 system

Amplifier:

The outputsignal from the electrode cap will be in therange of mV and μV. So, these signals will notbe suitable for signal processing. Hence thesesignals are fed to an amplifier. Each electrodeis connected to one input of a differentialamplifier (one amplifier per pair ofelectrodes); a common system reference electrodeis connected to the other input of each

9

differential amplifier. These amplifiers amplifythe voltage between the active electrode and thereference (typically 1,000–100,000 times, or 60–100 dB of voltage gain).

Analog to Digital Converter:

The output signals from the amplifier are analogin nature. They also contain some unwantedsignals. Hence the output signals are filteredusing high pass and low pass filters. The high-pass filter typically filters out slow artifactwhereas the low-pass filter filters out highfrequency artifacts. After the signal isfiltered they cannot be directly fed to adigital signal processors and microcontrollerunit as they are in analog form. Hence thesesignals are sent to an Analog to Digitalconverter to convert the incoming analog signalsin to digital signals.Signal Processor:

Using theoutput signal from the A/D converter, parameterextraction, pattern classification and patternidentification are done. Then the signals arefed to a Fast Fourier Transform Unit. This isdone to simplify the calculations. An FFT

10

algorithm computes the result in O(Nlog N)operations instead of O(N2) operations. Theoutput signals from the signal processor are fedto a Microcontroller unit.

Microcontroller unit:

Theoutput signals from the signal processor are fedto a microcontroller unit. This microcontrollerunit performs the robotic operation with thehelp of a stepper motor. It will control theoperations such as walking, running, etcdepending upon the input signal. For differentpatterns of input signals it will be pre-programmed to do a specific operation. Thereference signal will be already stored in themicrocon - troller memory in digital form.Usually an 8 bit or a 16 bit microcontroller ispreferred depending upon the number ofoperations to be performed. The complexity ofthe microcontroller programming increases withthe number of operations which has to beperformed.

11

Working of the Proposed System:

For every human activity the brain waves changesits pattern. For example, if a person moveshis/her hands then a specific pattern of brainwave is obtained and if the same person moveshis/her legs then a different pattern of brainwave is obtained. Even if a person thinks ofmoving his/her legs a brain wave of specificpattern is produced and it is sent to the legsand then the operation of moving the legs isperformed. The same brain waves are producedeven for a person who is not having his/herlegs. But the operation of moving the legs willnot be performed due to the absence of legs. So,just by thinking of moving the legs, a brainwave which is capable of performing a specificoperation is generated in the brain. Due to thelack of the appropriate system, the activitywill not be performed successfully. In theproposed system, the brain waves are pre-recorded for each operation to be performed andthese waves are used as reference signals. Thesesignals are stored in the microcontrollermemory. For each reference signal in the microcontrollermemory, the robotic leg is pre-programmed to doa specific operation. When the reference signal

12

matches with the actual signal from the user’sbrain, the robotic leg will do the pre-programmed operation with the help of themicrocontroller. For example, let us say thatthe user is thinking of walking. So a brain wavewill be produced. These waves are processed andthen it is converted in to digital signals.These signals are compared with the pre-recordedreference signals and a match in the signalpattern will be found in the microcontroller.The operation for this particular pre-recordedsignal will be pre-programmed in themicrocontroller circuit i.e. walking and thusthe microcontroller will send the control signalto the artificial robotic leg and the roboticleg will perform the required operation. Usuallya stepper motor controlled robotic leg is usedfor this purpose. Similarly to walking, other operations can alsobe performed using the artificial leg. Thissystem is very user friendly and the system canbe designed according to the user’s requirementsi.e. the number of operations required for theuser can be fixed by him and the system can bedesigned accordingly. So the number ofoperations that has to be performed by the legcan be increased or decreased and the complexityof the design varies accordingly. This idea can

13

be extended for both the legs and both the legscan be made to do operations like walk, run etcsimultaneously. Thus the system is versatile.This system is hundred percent feasible in thereal time environment and it can be implanted toany human irrespective of their age. AboveFigures shows some of the pictures of artificiallegs.

14

Fig:Proposed Model of the artificial legs fig: Internal appearance of the artificial leg

15

Fig 7: External Appearance of the artificial leg Fig8: Walking down the stairs using artificial legs

16

Fig 9: Artificial legs fitted to both the limbs Fig 10: Walking with artificial legs.

Power Supply:

Theseartificial legs are powered by a small lithium-ion battery which has to be charged once intoday. Lithium-ion batteries have very highcharge density (i.e. a light battery will storea lot of energy).They are of ultra-slim designand hence they occupy very less space. Moreovertheir life time will belonger when compared toother batteries. Hence they are preferred whencompared to other batteries.Moreover they havelonger life time when compared to otherbatteries.

17

Normal Artificial legs:

NormalArtificial Legs, available in the market, isvery costly. They use a group of sensors and acomplex algorithm for their operation whichmakes the existing system very costly. Thisdisadvantage hasbeen overcome in the BrainControlled Artificial Legs as they don’t use anysensors for their operation.Moreover the normalartificial legs are 100% dynamic in operation.Hence the chance of occurrence of anerror ismore in those systems. External appearance andoutput of both the legs are same. But themethodof operation is different. Hence the BrainControlled Artificial Legs are cost effective.

Difference between the Brain Controlled Artificial legs and the Normal Artificial Legs:BRAIN CONTROLLED ARTIFICIAL

LEGSNORMAL ARTIFICIAL LEGS

1. Ease of Construction 1. Complex in construction

Cost is not more thanRs.5,00,000

2. Cost is about $80,000-$90,000(Rs.35,00,000 to

18

Rs.40,00,000)3. User can have full control overthe artificial leg.

3. User cannot have full control over the artificial leg.

4. Semi-Automatic 4. Fully Automatic5. Sensors are absent. 5.Sensors are present6. Requires simple control unit.

6. Requires complex control unit.

Conclusion:

Forty years ago, thetechnology was so basic. Newton said. “Legsockets were made out of wood,offering theequivalent of a door hinge at the knee”. Butwith the recent advancement in thetechnology,Brain Controlled Artificial leg canbe made as a reality. The performance of theproposed systemwill bebetter than the existingartificial legs as the user has full controlover the Brain Controlled Artificial Legs.Henceit behaves like a normal human leg. The built-inbattery lasts anywhere from 25 to 40 hours soitcan support a full day’s activity. Therecharge can be performed overnight or whiletraveling in a car via acigarette lighteradapter.The cost of the proposed system is foundto be very less when compared to the existingones. So,

19

even the middle class people who cannot purchasethe existing artificial legs can make use ofthisproposedsystem. With this system life can bemade easier for the handicapped persons and theycan also dotheir day-to-day activities normallywithout any difficulties.

20