Study On Posture Control And Local Path Planning For Coal Mine Detection Robot

Rescue robot has become an important mean in the process of disaster succoring, but most rescue robots still rely on remote operation for motion control. Our research team of Xi'an University of Science & Technology proposed that the coal mine rescue robot robot was composed of carrier robot and detection robot, and the coal mine rescue robot could autonomous navigation and work. In this paper, the posture control technology and local path planning of coal mine detection robot in unknown environment are in-depth researched.In this paper, with analyzing the environment of coal mine accident and the mission of coal mine detection robot, the robot control system was designed. The distributed modularized architecture with three-layer was used for the control system, and the distributed hierarchical architecture was used for decision control system as the core of the robot control system. The principle of the control system was taken as the basis for building the mine robot control system experiment platform system. The control system can meet the requirements of detection capability in the coal mine of coal mine detection robot, and it has satisfactory performance in autonomy, openness and real-time.The motion mechanism of coal mine detection robot was designed as tracked robot with 4- parallel tracked limbs. The method was presented that the pose of equivalent robot which the non-holonomic constraints between parallel limbs of the detection robot and ground were eliminated through equivalent transformation of the motion mechanism was analyzed. By implementing this method, the detecting robot was equivalent into a 4-PPSR parallel robot, and the posture kinematic model of the moving platform in parallel robot was analyzed with the use of motion screw theory. With analyzing the static stability on the basis of the robot kinematics model and dynamic stability based on the changing analysis of potential energy, therefore, two conclusions were got respectively, one was static stability decision theorem based on projection stabilization cone through geometric analysis; another was the relationship between the speed and dynamitic stability, which was the important theory basis for keeping stable motion while the coal mine detection robot was driving. On the coordination for four basis behaviors of coal mine detection robot, driving behavior, attitude adjustment behavior, obstacle crossing behavior and path planning behavior, a reactive behavior coordinator system based on expert logic system for mine rescue robot was proposed. In the paper, the behavior pattern was analyzed of which in different road condition when robot was driving underground in coal mine. Among them, the attitude adjustment behavior on the operation of robot driving was emphasized researched. Therefore, based on the fuzzy control algorithm, the coal mine detection robot reactive behavior control for the attitude adjustment method was proposed, and with the experiment, it was testified that the control method was good adaptable for the complexity underground environment.In this paper, a local path planning method closer to the human mind, direction optimizing artificial potential field based on the rolling windows was proposed, which was for robot's path planning behavior of detecting mines in an unknown environment. The method used in classical artificial potential field algorithm which boils down the global information to a single field force was abandoned in the improved artificial potential field. The improved artificial potential field can resolve the best direction which field force is biggest among all driving direction for robot. The improved artificial potential field combining with the rolling windows path planning, the reaction local path planning method was presented. While the reactive path planning behavior of the robot could not achieve at the global target because of absence of global information, it will cause the non decision-making status and planning pitfalls and other issues probably. Therefore, the center average method was proposed of recalculate the maximum force to solve the non decision-making status. Foe solving the planning pitfall issues, it was required to deliberative planning and make judgments on the current situation. The deliberative planning process should analyze the travailed scenes based on the local grid-topological map which was analyzed through active growth Voronoi graph to generate the collision-free paths network consisting of Voronoi edges and nodes. The shortest escape path between escape target node and global goal node can be got through the Floyd shortest path searching method. The reaction- deliberating mixed local path planning method of coal mine detection robot was proposed, which was constituted by reactive path planning behavior and deliberative style planning process. By experiment, it was testified that the method can solve the obstacle avoidance problem effectively in the process of planning the trap and could be adapted to the complexity of coal unknown environment.Based on the theoretical research of this paper, the coal mine detection robot control system based on the study achievements of this paper was developed used in experimental robot which was developed by our research team through VC++. The experiments were operated in the laboratory environment and the simulated roadway environment with experimental robot, which indicated that the coal mine detection robot could achieve the understanding environment and path planning in an unknown environment, and could maintain the good posture in the condition of unconstructed roads. Above all, it is indicated the research of this paper provided with practicality and availability in the unknown complex environment.