Coal-based Multi-sensor Information Fusion Environment Detection And Risk Assessment

In recent years, the frequent occurrence of mine accidents bring great loss to life and fortune . Based on the complex environment of the mine ,we designed a multi-sensor information acquisition and fuse system. Meanwhile, a new Rough Set-Neutral Network algorithm for mine risk-assessment is put up to make accurate real-time assessment of the mine environment.The system uses Freescale's latest 16-bit MCU MC9S12DG128B as core to compose an intelligent control node, Through which we communicate with the sensors for information acquisition. In order to simplify the system architecture and improve the reliability, CAN bus is used as communication structure of the intelligent node system. This ensure the real-time and accuracy of the dangerous environment detection in hardware.In control algorithms, neural networks has the advantage of parallel processing and information storage. While Rough set can eliminate the noise and redundant target samples by attribute reduction and value reduction to the data . Based on this,the article combine the two theory into Rough Set -Neutral Network (RSN), Explained it's principles, algorithms and implementation process .Then uses it to realize the danger assessment of the mine environment.The combination of the two method can not only reduce the size of the network ,then reduce network burden of training and study.but also can improve the predict accuracy of the neutral network. Simulation results show that the RSN can accurately assess the danger of the mine environment, meanwhile has a strong anti-jamming capability and a high accuracy, thus verifying the effectiveness andfeasibility of the system.