一种多关节履带式机器人自主越障运动规划

Autonomous motion planning for obstacles-climbing of a multi-track robot

摘要:

为提高矿难搜索机器人的自主运动能力,建立了其翻越垂直障碍的运动学模型,并且对其翻越垂直障碍的能力进行了分析,提出了一种基于嵌套人工鱼群算法的机器人自主越障运动规划方法.采用三层四组人工鱼群算法削弱4个自变量之间的耦合作用,并且利用一种快速随机搜索方法缩小关节角度的搜索范围,从而实现了满足关节角度变化最小这一优化指标的机器人越障逆运动学求解的要求.MATLAB的仿真实验证明:该方法相对传统算法可以利用更少的人工鱼数目在较少的代数内有效地得到机器人越障中每一步骤的最优关节角度值,从而为机器人的控制提供依据.利用仿真结果对机器人进行了双单元模式的自主越障运动规划实验,实验结果表明机器人能够顺利实现越障.

Abstract:

To improve the autonomous ability,the obstacles-climbing module of a searching robot under mine disasters is built.Based on the model,the ability of obstacles-climbing robot is analyzed and an nested artificial fish-swarm algorithm(NAFSA) is developed to solve the automatic motion planning of the robot.The method adopts the AFAS in three layers of four groups which can weaken the coupling of the four variables,and uses the rapidly random exploring method for reducing the variables bounds to meet with the minimum changes of joint angles.Then,simulations in MATLAB show that this method can get the key joint angles more effectively and in less intervals than traditional method,and provide reference for the robot control.Finally,the obstacle-climbing experiment of double units mode is carried out,and the simulation result verifies that the robot can climb over the obstacle.