摘要
针对未知环境清扫机器人的环境识别和建模,提出了一种基于BP神经网络的环境建模算法。首先提出了一种利于计算机识别和处理的室内环境拓扑结构,再利用3个超声波传感器获得周围物体的距离信息,对这些距离数据进行归一化处理之后利用BP神经网络来识别出物体的形状,完成环境的建模。给出了系统的组成结构、室内环境拓扑结构、BP神经网络原理及学习步骤。并对实验获得的一组数据进行了仿真,结果表明这种算法对于环境识别和建模是一种有效的方法。
Aimed at the cleaning robots＇ environment identify and modeling in unknown environment, a kind of environment modeling arithmetic based on the BP neural network was given. Firstly, a kind of benefit for computer identifying and processing of indoor environment topological structure was put forward, then three ultrasonic sensors were used to acquire the distance information of surrounding objects. After dealing with these data, BP neural network was used to identify the shape of objects and to finish environment modeling. The constituting structure of the system, the topological structure of the indoor environment, the principle of BP neural network and the step of studying were given. The data received in the experiment were simulated and the result shows that it is a valid method to identify and model the environments.
出处
《机床与液压》
北大核心
2007年第9期93-96,共4页
Machine Tool & Hydraulics
