Research Of Multi-Sensory Information System Of Intelligent Myoelectric Prosthetic Hand

The purpose of the research on multi-sensory prosthetic hand is to improve theself-care ability of the disabled and narrow the difference between the healthy and thedisabled, supporting the physical disabilities’mental health and promoting the socialcivilization and the development of medical benefits. At the same time, the artificialprosthetic hand is cross-disciplinary fields of multi-sensor technology, informationfusion technology, biomedical engineering, and electronics information technology,promoting mutual penetration of today’s cutting-edge technologies. The upper limb isvery complicated, having rich sensory feedback. However, the current commercialelectric prosthetic hand controlled by surface electromyography (SEMG) signals doesnot have information perception feedback of grasping objects. In order to makeprosthetic hands acting like human hands as much as possible and improve thebiomimetics and intelligence, it is necessary to add multi-sensor perception system onelectric prosthetic hand system.The handicapped adorned with prosthetic hand take the SEMG of the stump asthe control signal to achieve natural motion control, the process involves the SEMGsignal processing, motion recognition and motion control technology. Thecombination of multi-sensor perception and SEMG electric prosthetic hand will makeprosthetic hand more bionic and intelligent.Therefore, this article takes theoretical and experimental exploration onproximity detection, hardness detection and thermal conductivity measurements whengrasping objects after the SEMG signal processing and recognition. The authordesigns and develops the multi-sensor perception hardware circuit, eventuallyintegrated into the artificial prosthetic hand system. The research and innovations arefollows.(1) Literatures at home and abroad of prosthetic hand technologies and kinds ofnew type of sensors integrated into prosthetic hand are reviewed. The researchbackground, status and significance of hardness sensors, proximity sensors, thermalconductivity sensors and prosthetic hand are briefly generalized.(2) In order to eliminate the signal coupling cross-talk among the acquiredmulti-channel SEMG, a method combines the wavelet analysis and the blind source separation (BSS) is proposed, improving the signal-to-noise ratio. The method oftime-frequency combination feature extraction and fusion is adopted in order to fusionthe ability differences of different BP network in pattern recognition, making full useof the complementary of every feature, improving the recognition rate of motionmode and the robustness of the prosthetic hand.(3) Based on the designed robotic tactile sensor and the control speed of paws, abiomimetic hardness sensor suitable for electric prosthetic hand is designed. A simpleand practical method of testing grasping objects’hardness is proposed. The resultsindicate that with the same clamping speed and time interval, the cumulative value ofthe pressure signal expresses the hardness’efficiently.(4) A thermal conductivity sensor installed on the fingers of prosthetic hands isdeveloped and its corresponding signal processing circuits is designed. Forenvironment and climate variability, a humanoid PWM temperature control method isproposed, avoiding the error of the detection system caused by the workenvironment’s inconsistency. The designed thermal resistor array detection unitlessens the blind detection area efficiently, reduces the requirements of the contactposition and increases the stability and sensitivity of the perception system.(5) A infrared proximity sensor installed on the fingers of prosthetic hands isdeveloped, the proximity signal is launched by the modulation circuit anddemodulated by the demodulation circuit. The detection system can judge the degreeof closeness accurately and rapidly, and provide feedback to controller to adjust speedof driver motor so as to grab objects softly.(6) The multi-sensors are integrated in prosthetic hands. Some typical objects areclassified through the binary decision tree information fusion algorithm. The researchresult provides technical routes and methods for development of SEMG electricprosthetic hands with multi-feelings.