Forecast of Gas Distribution Area for Coal Mine Robot

Abstract

Abstract. Limited by the explosion-proof safety level, the coal mine gas robot needs to avoid the dangerous area of gas. In this paper, a hybrid particle swarm optimization algorithm combined gray neural network of the gas distribution prediction method are proposed to get more accurate prediction of the gas distribution ahead 10m of coal mine robot, considered the main factors that affect the concentration of gas, such as temperature, wind speed and direction, distance and so on. Experimental results show that the overall prediction accuracy of the HPSO-GNN prediction method is improved compared to the GNN prediction method. The method can accurately predict the gas concentration distribution area, and provides a basis for the coal mine robot to avoid the dangerous area of gas, and ensure the safety operation of the coal mine robot in the coal mine.