基于遗传算法的多边形分割AUV全局路径规划

Genetic algorithm based on polygon segmentation AUV global path planning

针对遗传算法在实际AUV全局路径规划应用中出现运算数据大、路径规划有尖峰等问题,提出了新型路径规划方法。利用平面直角坐标系实现环境的建模,将障碍物简化成多边形并分割为三角形。路径用首尾相接的线段表示,通过固定横坐标,随机生成纵坐标的方式实现遗传算法二进制编码,对障碍物三角形交叉判断,路径距离运算实现适用度函数编写。对遗传之后的路径通过避障、删除节点、平滑的操作确定最终优化路径。结果表明,对障碍物的三角形简化实现了在遗传操作中的程序优化,利用避障、删除多余节点、平滑操作实可很好的消除尖峰,可寻找一条相对较优的路径。

Appear in the actual AUV global path planning applications for genetic algorithm computed data, path planning issues such as spikes, a new path planning method is proposed. Plane rectangular coordinate system environment modeling obstacle simplified polygon is divided into triangles. End-to-end path segment, the fixed abscissa, ordinate randomly generated binary-co- ded genetic algorithm obstacle triangle intersection determination, the path from the operation prepared by the applicable func tion. After genetic path through the obstacle avoidance, delete nodes, smooth operation to determine the final optimization path. Results show that the simplified implementation obstacles triangle program optimization in genetic operation, use of obstacle avoidance, remove redundant node, smooth operation, can very good eliminate peak, can find a relatively optimal path.