摘要
传统基于单目视觉机器人避障方法在复杂变化环境中存在的测距精度低以及稳定差的弊端,机器人自主避障性能差,因此,提出基于神经网络的机器人激光测距方法,分析信息融合方法在机器人激光测距中的运用过程,设计机器人运动环境包括8个不同级别,将激光传感器在小车身上的8个安装位置,塑造基于T—S模型的模糊神经网络,运算出神经网络各层节点函数,并对神经网络进行训练。设计基于T—S模型模糊神经网络的机器实施激光测距的控制模型结构,运用启发式推导的模糊规则协助机器人的导航调控功能,实现移动机器人激光测距运动,准确避开障碍物。实验结果说明,所提方法避障效果佳,具有较高测距精度和稳定性,提高机器人自主避障性能。
The traditional monocular vision robot obstacle avoidance method has the disadvantages of low ranging accuracy and poor stability in complex changing environments,and the robot autonomous obstacle avoidance performance is poor. Therefore,a neural network based robot laser ranging method is proposed. The application process of information fusion method in the robot laser ranging is analyzed. The robot motion environment is designed including 8 different levels. The laser sensor is mounted on the small body with 8 mounting positions,and the fuzzy neural network based on T-S model is constructed to calculate the nodes of the neural network and train the neural networks. The control model structure of laser ranging based on T-S model fuzzy neural network is designed. The heuristic-derived fuzzy rules are used to assist the robot＇s navigation and control functions to realize the laser ranging motion of mobile robots and avoid obstacles accurately. The experimental results show that the proposed method has good obstacle avoidance effect,high ranging accuracy and stability,and improves the robot＇s autonomous obstacle avoidance performance.
作者
郑志材
ZHENG Zhicai(Guangdong College of Business and Technology, Zhaoqing Guangdong 526040, China)
出处
《激光杂志》
北大核心
2018年第11期167-172,共6页
Laser Journal
关键词
神经网络
机器人
激光
测距方法
T—S模型
信息融合
neural network
robot
laser
distance measuring method
T - S model
information fusion
