Intelligent Fusion Method Based on BP Neural Network for Robot Suit Pressure Sensing Array Data

摘要

In order to eliminate the crosstalk influence existed among temperature, voltagefluctuation and sensor signal of robot tactile sensor array unit, the paper presented a sort ofinformation fusion method in large scale sensor array based on BP neural network.By means of learning training with weight for neural network, the method can effectivelyeliminate the crosstalk influence for output characteristics of pressure sensing array sensoramong non-target parameters and large scale sensor signals such as the environmenttemperature, voltage disturbance and so on, and thereby it improves the stability andreliability of robot tactile sensing suit system. Laboratory tests demonstrated that the errorof the suit pressure sensor array data is less than 5%. The experimental results show that theintelligent fusion method presented in this paper can be accepted in engineering application,and the method is be propitious to improve intelligent judgment and information utilizationratio of robot system.