基于蚁群势场算法的建筑材料运输机器人智能避障方法

100124)2(BeijingChao-YangHospital,CapitalMedicalUniversity,Departmentof

摘要:

针对现有建筑材料运输机器人避障中存在的全局寻优能力差,易与移动障碍物发展碰撞的不足,设计了一种蚁群势场算法。首先分析了蚁群算法下蚂蚁个体信息素浓度的累积过程,通过构建人工势场求解引力和斥力的合作,将其作为优选蚁群算法启发因子的重要约束条件；其次引入SA算法对蚁群势场算法做二次优化,将降温的过程视为一个全局优化的过程；最后在局部避碰方面构建了质量点模型,通过评估机器人当前位置、运行速度和障碍物位置等信息建立惩罚函数,并将惩罚函数值降至最低,避免出现与障碍物的碰撞。实验结果显示:提出算法有更高的迭代效率,复杂动态条件下最短行进距离和时间分别为110.6m和115.1s,且在局部未出现与其他移动机器人的碰撞情况。

Abstract:

An ant colony potential field algorithm was designed to solve the problem of poor global optimization ability and easy collision with moving obstacles in existing building material transport robots. Firstly, the accumulation process of individual pheromone concentration under ant colony algorithm is analyzed, and the cooperation of gravity and repulsion force is solved by constructing artificial potential field, which is regarded as an important constraint for selecting the heuristic factor of ant colony algorithm. Secondly, SA algorithm is introduced to optimize the ant colony potential field algorithm twice, and the cooling process is regarded as a global optimization process. Finally, a mass point model is constructed in terms of local collision avoidance, and a penalty function is established by evaluating the robot"s current position, running speed and obstacle position, and the penalty function value is reduced to the minimum to avoid collision with obstacles. The experimental results show that the proposed algorithm has higher iterative efficiency, and the shortest travel distance and time are 110.6m and 115.1s respectively under complex dynamic conditions, and there is no local collision with other mobile robots.