Development Of The Coal Mine Disasters Information Detection Robot And Map Building And Path Planning Research

Information detection on the coalmine disaster scene has been the research focus for experts home and abroad.After disaster,the earlier rescue action performed,the more lives and property would be recovered,but the precondition is to find whether the environmental information on the spot can meet the requirement of rescue action.Thus,coal mine information detection robots,who can replace the rescue staff in accomplishing the task of information collection on the spot,will be a good solution on the issue.The author of this paper,aiming at the research of 3D environment map and path planning on coalmine disaster scene,has developed a two-level coalmine disaster information detection robot.Upon the present research on coalmine information detection robots,the paper presents the two-level coalmine information detection robot system based on the haven and double T-steel track of underground coalmine shaft.When disaster occurs,first level robot,starting from the haven,travels along double T-steel on the roadway roof;it will release the second level robot it carries to detect disaster information further when it can not proceed.The paper presents a communication system linking the coalmine disaster scene and the ground.The system is based mainly on the wireless technology of WIFI and the protocol TCP/IP for its web construction.The system has some improvement on RSSI path damage model,then a path damage model with multi-path effect suitable for coalmine shaft is produced,upon which the wireless nodes are mounted in coalmine shaft simulation lab,the results show that communication signals are clear and stable.For the second level information detection robot,the paper analyzes its static stability and concludes the position relationship between the center of robot and the geometrical center.The paper,aiming at the robot's ability in climbing over obstacles,constructs a simplified model of the robot and concludes the theoretical height it may climb over and the biggest slope it may climb by analyzing the influence on the robot's climbing over obstacles by the robot's center change brought about by rocker arms' revolving around the axis.The paper,focusing on the auto navigation of the coalmine information detection robot,proposes a map construction calculation method by employing modularized local raster map to indicate the complex 3D environment.To solve the problem of the difficulty in obtaining the 3D environment information,the paper presents the application of TOF camera to collect the information and the formation of cloud coordinates matrix for 3D dots.The paper explains how to locate the coordinate dots on the local raster map by the calculation of the dots of the matrix,how to construct the whole local raster map by updating the occupation probability of raster unit,how to simplify the calculation of path planning by identifying the passage area from non passage area on the basis of the division of types of obstacles according to the density of coordinates in the raster units.In the paper,to fix the inadequacy of path planning calculation in the traditional raster environment,a two level A-path planning calculation method is added in raster obstacle occupation probability.Under traditional A-method in a large scale raster environment,it would take too long time to calculate.To overcome this,this paper proposes there are two level raster maps: one is higher level raster map;the other is lower level raster map.The higher level raster map decides the searching direction while the lower level raster map is used for the local path planning.To make up the inadequacy in raster map's obstacle indication,the raster occupation probability function is added in the calculation method,thus the prediction accuracy of calculation on the path length is improved.The best path is achieved by the combination of higher level and lower level calculation method.Aiming at solving the problem in path planning of coalmine disaster information detection robot's auto navigation,the paper proposes an improved mark modifying calculation method.The method can solve the problem of multi-target path planning in complicated environment.The paper presents a principle of cumulative cost and the smallest mark sequencing to solve the problem of mark sequencing of the method,and testifies that the solution brought about by the principle is Pareto best path.For the quality of the primary solution,the paper sets two judging standards so that invalid solution,which can not lead to Pareto best path,will be deleted at the early stage of calculation method.In applying the improved mark modifying calculation method,itwill take too long time to undergo multi-target optimization algorithm.Thus the paper presents a mark modifying calculation method with weight coefficient added,so that the multi-target optimization becomes a single target optimization by adding weight coefficient to various effects.The simulation experiment shows it takes fairly short time to obtain the Pareto best path by applying the method in large scale raster map,and optimized paths with various properties could be obtained according to the requirement of decision makers.