摘要
针对基于生物激励神经网络的清洁机器人遍历路径规划算法的遍历面积重复率和遍历路径总长度均较大的问题,对该算法进行了改进:在脱困算法中,采用实时监测机器人邻域神经元状态的方法,使机器人脱困路径缩短;引入邻域神经元状态准则,使机器人在遇到孤岛障碍物避障时,先沿障碍物边沿遍历.仿真结果表明,改进算法可以有效降低遍历面积重复率、遍历路径总长度和转弯次数.
In view that the traversal area repetition rate and the total length of the traversal path of the traversal path planning algorithm of the cleaning robot based on the biologically inspired neural network are large, thealgorithm was improved. In the relief algorithm, the method of real-time monitoring of the neurons in the neighborhood of the robot was adopted to shorten the path for the robot to get out of difficulty. The state criteriaof neighboring neurons were introduced to make the robot traverse along the edge of the obstacle traversal when obstructing obstacles in an island. Simulation results showed that the improved algorithm could effectivelyreduce the traversal area repetition rate, the total length of the traversal path and the number of turning.
作者
张志远
赵幸
靳晔
ZHANG Zhiyuan;ZHAO Xing;JIN Ye(College of Mechanical and Electrical Engineering,Zhengzhou University of Light Industu,Zhengzhou 450002,Chin)
出处
《轻工学报》
CAS
2018年第4期73-78,85,共7页
Journal of Light Industry
基金
河南省重点科技攻关项目(17210210057)
郑州轻工业学院星空众创空间孵化项目
关键词
清洁机器人
路径规划
生物激励神经网络
cleaning robot
path planning
biologically inspired neural network
