Underground location algorithm based on random forest and environmental factor compensation

Abstract Aiming at the poor location accuracy caused by harsh and complex underground environment, long strip roadway, limited wireless transmission sparse anchor nodes, an algorithm based on random forest compensation for environmental factors was proposed. Firstly, access point (AP) network model tunnel environment were analyzed, fingerprint built. And then Received Signal Strength (RSS) analyzed Kalman Filter in offline sampling real-time positioning stage. Meanwhile, target speed constraint condition introduced to reduce error factors. The experimental results show that proposed solves problem of insufficient large fluctuation affected when nodes are sparse. At same time, average reaches three meters, which can satisfy application rescue, activity track playback, disaster monitoring positioning. It has high value environment.