Multi-sensor data fusion method based on spiking neural network

Abstract

The invention provides a multi-sensor data fusion method based on a pulse neural network. The method comprises the following steps: converting multi-sensor data into a pulse sequence through pulse frequency coding; extracting basic features of each sensor by using a pulse neural network, and outputting a new pulse sequence; performing cross-sensor time sequence alignment and aggregation on the new pulse sequence by using time information of pulses, fusing features of different sensors, and generating a cross-sensor comprehensive feature representation pulse sequence; inputting the comprehensive feature representation pulse sequence into a pulse neural network with set weight parameters; extracting advanced features related to the control task; and outputting a decision signal and a control signal according to a processing result. According to the invention, multi-sensor data are fused through the pulse neural network, and the method has the advantages of low power consumption, high real-time performance, high accuracy and high robustness; through the multi-hidden-layer structure of the spiking neural network, multi-sensor data are extracted and fused more sufficiently, and the reliability and accuracy of the system are improved.

Images