Research On Autonomous Obstacle Avoidance System For Unmanned Vehicle In Environmental Reconnaissance

It has become a trend to use unmanned environmental reconnaissance vehicles instead of combatants to perform reconnaissance missions in a variety of unknown and dangerous situations.This article selects the research background as the obstacle avoidance algorithm for battlefield environmental reconnaissance unmanned vehicles.After comparing and analyzing the characteristics of the battlefield environment,the artificial potential field method was selected as the basic obstacle avoidance algorithm for the unmanned environmental reconnaissance vehicle.After analyzing the common problems of obstacle avoidance,the problems existing in the practical use of the algorithm were obtained.For the point problem,a“meter” coordinate trial-and-error method based on a conjugate gradient search was proposed to solve the local minimum point problem.For the unreachable problem of the target point,an artificial potential was redefined in the area affected by obstacles near the target point.The method of the repulsive force field function of the field method is used to solve the problem of unreachable target points.In order to improve the replaceability of the sensor,eight 90 °single-line laser sensors are selected to form a circular array,which ensures that the scanning range can reach 360 ° in the case of individual sensor failure.The related circuit is designed based on the delay measurement chip TDC-GP2,and the actual measurement compares the accuracy of the measurement data of the single-line laser sensor and the lidar.Based on the Mobot Sim software environment,the algorithm simulation was carried out.Under the conditions of known,unknown and semi-unknown obstacles,the improved artificial potential field method was simulated and evaluated to verify the effectiveness of the algorithm.An obstacle processing method based on a single-line laser sensor array and an obstacle environment map processing method are proposed.The GPS multi-point differential positioning is used in an unknown environment to realize the real-time localization of an unmanned environmental reconnaissance vehicle in the environment and calculate other obstacles.The absolute coordinates are substituted into the obstacle avoidance algorithm to plan the path,and the absolute coordinates of the known obstacles are calculated in a known environment,and compared with the known relative coordinate data,the effectiveness of the GPS error correction method is verified.An unmanned environmental reconnaissance vehicle test system based on a single-line laser sensor array was built.The effectiveness of the improved method of artificial potential field was verified by simulating local minimum point problems and unreachable target points in the test environment.Under the same environmental conditions,the comparative experiments of obstacle avoidance performance of fuzzy logic algorithm,genetic algorithm,artificial potential field method and improved artificial potential field method were performed respectively,and the effectiveness of the improved artificial potential field method was verified.The simulation and experiment of the improved artificial potential field method are performed under semi-unknown environmental conditions,and the experiments verify the effectiveness of the algorithm.