摘要
针对移动机器人避障的特点,提出了一种基于神经网络的动态避障控制方法。介绍了避障行为的决策、基于神经网络的机器人在避障过程中的运动控制等。该方法不用考虑障碍物的运动状态,简化了机器人避障的步骤,机器人能够根据各种情况灵活地判断是否避障以及灵活地选择适当的避障方式,提高了机器人避障的灵活性和鲁棒性。仿真试验证明这种方法是可行而有效的。
According to the characteristics of the mobile robot in obstacle avoidance, a method of dynamic obstacle avoidance based on neural network is presented. This article presents the decision of obstacle avoidance behavior, robot moving control based on neural network during obstacle avoidance process. This method does not necessarily take the velocity and the direction of the obstacles into account, so predigests the procedure of obstacle avoidance. The robot can judge flexibly whether it needs to avoid the obstacle and can choose proper ways of obstacle avoidance according to diversified circumstances. It improves flexibly and adaptively the robot obstacle avoidance. The experiment of simulation indicates that the method is feasible and valid.
出处
《农业装备与车辆工程》
2006年第8期22-24,共3页
Agricultural Equipment & Vehicle Engineering
关键词
移动机器人
神经网络
避障
mobile robot
neural network
obstacle avoidance
