基于量子蚁群算法的机械手臂避障路径规划

Mechanical Arm Avoidance Path-Planning Based on QACA

提出在障碍空间下利用量子蚁群算法求取6R机械手臂逆解的一种通用解法。首先根据机械手臂各关节变量的DH参数建立以关节变量为自变量的目标优化函数F,并且引入层次包围盒OBB的碰撞因子以及可操作空间的约束,再建立基于Bloch球面的三维量子旋转搜索空间,利用量子蚁群算法搜索出F函数取到最大值时所对应的关节变量,即机械手臂的逆解,因此将机械手臂的逆解问题转化为基于关节变量的多元函数求极值问题。鉴于基本蚁群算法在优化之初可能存在搜索缓慢、蚂蚁数量限制优化解的范围、以及算法可能陷入局部极值等问题,采用量子蚁群算法利用量子计算的并行性、三链编码机制以及量子旋转门、非门对蚁群算法进行改进。改进后的量子蚁群算法与基本蚁群算法,通过MATLAB Robotic Toolbox工具箱对6R机械手臂进行运动学仿真与CCS DSP联合开发实验,结果表明量子蚁群算法在优化F函数中能够快速收敛,降低了对蚂蚁种群数量的依赖,并且扩大了空间解的范围,从而证明方法可行。

A general method for applying quantum ant colony algorithm to obtain the inverse solution of 6R mechanical arm is proposed.Firstly, according to the DH parameters, establishing optimization function F with the aim joint variables as independent variables, joining the OBB bounding box collision factor as well as the operational space constraints.Secondly, establishing three-dimensional quantum rotation search space based on Bloch sphere and using quantum ant colony optimization algorithm to search out the corresponding joint variable when the function F is at its maximum value, which is the inverse of the robotic arm.Therefore based on the inverse solution of mechanical arm problem can be converted to extremum problem about joint variables of multivariate function.Given the optimization search is slowed at the beginning of the basic ant colony algorithm and ants limit range is small and the algorithm of the optimal solution may be trapped in local minimum problem, this paper take advantage of parallelism computing from the quantum ant colony algorithm and three chain encoding mechanism.In addition, a quantum revolving door and NOT gate also are used improve the ant colony algorithm.Finally, the improved quantum ant colony algorithm and the basic ant colony algorithm were performed by MATLAB Robotic Toolbox to perform kinematics simulation and development experiment with CCS DSP on 6R mechanical arm.The results showed that: the quantum ant colony algorithm can converge in the optimal F function rapidly.What is more, it reduces dependency on ant population and expands the scope of spatial solution, which expects that this method is feasible.