Research of environmental modeling method of coal mine rescue snake robot based on information fusion

摘要

A coal mine rescue snake robot with four orthogonal joints is developed. Taking the robot as an experimental platform, aiming at how to model and identify the environment of the coal mine tunnel after the disaster, a multi sensor data fusion algorithm based on genetic algorithm optimization of the variable structure fuzzy neural network (GAFNN) is proposed. Firstly, the fuzzy neural network model is structured by combining fuzzy framework with neural network. Secondly, the distance information between the snake robot and obstacle can be obtained by using ultrasonic sensors and infrared sensor, and the azimuth information of the target location can be obtained by night vision device. The information is fused by the fuzzy neural network. The neural network is introduced into the inference of fuzzy rule. Due to the fuzzy rule base is adjusted by variable structure, the self-learning of fuzzy membership functions and the preferential extraction of fuzzy rules are realized. Finally, the genetic algorithm is used to optimize the learning parameters of the variable structure fuzzy neural network, and the fast and global optimization of the parameters is realized. It overcomes the shortcoming that the conventional BP algorithm is easy to fall into local optimum, and improves the optimization performance. So that the error between the actual output and the expected output of the network is minimized, and the high accuracy identifying of eight kinds of typical environmental models which are established is realized. By comparing the simulation results of environmental models identification, the GAFNN algorithm can get higher identifying accuracy than the non-optimized FNN algorithm. The GAFNN algorithm is applied to the function of obstacle avoidance of developed snake robot, and it lays the foundation for further path planning.