Recognition Technology, Based On Information Fusion Of Multi-sensor Remote Control Lhd Environment

The working environment is not only impacted by natural factors, such as dust, noise, high temperature, when Remote Control Underground LHD works in the tunnel. Both the mining ores and other bad scheduled LHD also have an impact on it. In recent years, the remote control technologies are successively adopted by manufacturers of LHD at home and abroad, so that the operators are outside of danger. The important premise of remote control is obtaining accuracy and real-time environment information of LHD.Moreover, to achieve obstacle avoidance, safe driving, navigation or positioning control, path planning and clusters optimization of Remote Control Underground LHD, we must acquire its working roadway information in real time. The central control room will give the right instructions, while accessing the shape, size and location characteristics of obstacles on current environment.For the realistic value and significant reason, this paper gives a preliminary study on environmental recognition technology on LHD. There are some limitations in making use of camera, because dust and smoke always exist in the roadway. Therefore, we exclusively concern on the situation, in which with single type sensor. So in this paper, we relay on multi-sensor data fusion, ultrasonic distance measuring and convex shell technology to build the environmental recognition system. The main contents are as follows:1. The number and distribution location of environmental recognition module when we extract the characteristics of current environment objects. An environmental recognition module is composed of an ultrasonic sensor and an actuator. Each ultrasonic distance measuring module completes environment scanning at a range of 0 to 180 degree in a cycle, which is fixed on an actuator. In order to obtain the environment redundant information greatly, we choose 5 recognition modules and set them in staggered arrangement.2. The time sharing switches control of environmental recognition module. Through PIC single chip to select which recognition module is going to work. As soon as we assured the one, PIC single chip will give the instruction"Starting to 16 bit distance measure order".3. A new centralized multi-sensor weighted fusion algorithm is put forward. There might be several obstacles in the roadway at the same time, so the algorithm is mainly used to get the information about the nearest obstacle, which in front of Remote Control Underground LHD, in order to analyse the characteristics of the target object.4. To finish the extraction of two-dimensional outline characteristics about target object. Making use of fast convex shell algorithm to sort the points and kick out concave points. Finally the outline and size of target obstacles are given.