Outburst is one of the main types of gas disaster; this paper focuses oncoal and gas outburst prediction. Based on the study of the mechanism andcharacteristics, Building the model of multi-sensor information fusion coaland gas outburst prediction loop system and apply to coal and gas outbustprediction. Increasing the feedback link in general framework ofmulti-sensor information fusion, and giving a closed loop fusion model onforecast of gas outburst. And using sensor management subsystem asfeedback loops to constitute a closed-loop mode. Based on the comparativeanalysis of the various algorithms, BP neural network is selected as thecharacteristic layer fusion method. Because of the disadvantage of B-Pneural networks, Further proposed the D-S evidence theory as adecision-level fusion method. Using B-P neural network output and severaltypical indexes as the D-S evidence theory’s evidence for decision levelfusion. Finally this paper constitutes the characteristic layer fusion anddecision-making level hierarchical structure.fou increasing the reliability ofdecision-making. This paper selected some typical high gas outstandingmining’s data to validate the proposed multi-sensor information fusion system model of gas outburst forecast. Experimental results show that thisprogram is well to meet the requirements of the actual forecast in practicalapplications. |