面向溺水救援机器人平稳跟踪的模糊比例微分控制视觉伺服方案

Fuzzy Proportional Plus Derivative Control Visual Servo Scheme for Stationary Tracking of Drowning Rescue Robots

针对溺水救援机器人的平稳跟踪需求,提出一种基于模糊比例微分控制的视觉伺服方案,通过场馆顶部安装的俯视摄像头获取机器人的坐标位置,使用虚拟导航线设计动态行走路径,计算偏移方向、偏移角并制定转弯规则;利用三角隶属度函数制定模糊比例、微分控制规则表,根据2种规则表中的比例、微分系数分别控制机器人的偏移角、距离偏移量及其变化率,实现机器人的跟踪与避障的平稳性调整;机器人采用核相关滤波跟踪算法对溺水人员进行跟踪,在溺水人员位置变化时调整运动方向,依靠实时、稳定的声呐信息设置偏移角,实现避障功能;对所提出的方案在先锋P3DX型机器人上进行实验验证。结果表明:机器人能及时应对目标变化,没有跟错、跟丢现象;在不同工况时的横滚角均小于1°,避障率达到100%,能够快速、稳定、准确地到达目标位置,达到溺水救援机器人的前期应用要求。

To meet stationary tracking requirements of drowning rescue robots,a visual servo drowning rescue scheme based on fuzzy proportional plus derivative control was proposed.Coordinate positions of the robot were obtained through an overhead camera installed on the top of the venue,and a dynamic walking path was designed by using virtual naviga-tion lines.Offset direction and angle were calculated,and turning rules were formulated.A triangular membership func-tion was used to develop fuzzy proportional and derivative control rule tables.Offset angle and distance offset as well as their change rates of the robot were controlled according to proportional and derivative coefficients in the two rule tables to complete tracking and obstacle avoidance smoothness adjustment of the robot.The robot adopted kernel correlation filtering tracking algorithm to track the drowning person,and adjusted motion direction of the robot when position of the drowning person changed.The robot relied on real-time and stable sonar information to set the offset angle and achieved obstacle avoidance function.Experimental verification of the proposed scheme was conducted on Pioneer P3DX robot.The results show that the robot can respond to the change of the target in time without phenomena of mistracking and losing.All roll angles under different working conditions are less than 1°,and the obstacle avoidance rate reaches 100%.The robot can reach the target position quickly,stably,and accurately,which meets early application requirements of drowning rescue robots.