摘要
传统路径规划算法针对多目标情况,主要依据多单一信息融合结果选择避障路径,在中规模的污泥纠缠区域中容易陷入盲区,无法对污泥纠缠环境下的机器人路径进行准确的规划。为此提出一种改进的机器人视觉纠缠摆脱路径规划方法,借助机器人视觉仪器采集污泥纠缠特征,用归一化方法把视觉信息融入到规划模型中进行最佳路径的选择,将机器人摆脱污泥纠缠以及最短路径的要求融合成一个适应度函数,通过遗传算法搜索获取最佳机器人摆脱路径。实验结果说明,该方法对于污泥纠缠环境下机器人摆脱路径规划长度以及效率都优于传统模型,具有较高的鲁棒性。
Traditional path planning algorithm for multi-objective situation mainly depends on the results of single information fusion to choose the obstacle avoidance path. In the mesoscale sludge entanglement area, it is easy to catch in a blind area and unable to make accurate planning of the robot path under the environment of sludge entanglement. An improved path planning method t based on robot visual algorithm is proposed. The robot visual instrument and normalization method are applied to make an optimal path choice. Then, the requirements of which the robot can get rid of the sludge and find the shortest path. Finally, the genetic algorithm is applied to the search of the best robot path. The experimental results show that the method can get rid of sludge entanglement, and the length of the path planning and efficiency is better than the traditional model, and it also has better robustness.
出处
《控制工程》
CSCD
北大核心
2014年第3期365-368,共4页
Control Engineering of China
基金
国家自然科学基金项目(61203136)
关键词
污泥纠缠
机器人
神经网络
遗传算法
sludge entanglement
robot
neural network
genetic algorithm
