基于混合势场的移动机器人视觉轨迹规划

Visual trajectory planning for mobile robots based on hybrid artificial potential field

针对载有单目相机的非完整移动机器人,提出基于混合人工势场的能够满足非完整约束的路径规划方法,能够解决移动机器人运行过程中的视野约束问题.在传统人工势场的基础上,提出混合人工势场的概念,势场中一部分是只对角速度起作用的目标点偏转势场和视野约束排斥势场,另一部分是同时影响线速度和角速度的目标位姿吸引势场;其中视野约束排斥势场定义在图像空间内;目标位姿吸引势场和目标点偏转势场定义在笛卡尔空间,利用单应性矩阵三维重建的结果构造势场函数.在混合人工势场的控制下,机器人能够在同时满足视野约束和无侧滑约束的条件下平滑地移动到目标位姿.仿真结果证明了该方法的有效性.

A visual trajectory planning method was proposed based on hybrid artificial potential field for anonholonomic mobile robot equipped with a monocular camera in consideration of field-〇f-view （F O V ） constraints.A hybrid model was designed based on the concept of conventional artificial potential field. Themodel consisted of two parts. One part only affects the angular velocity and involves the terms for targetdeflection and field of view constraints, and the other part affects both the linear velocity and the angularvelocity. The potential field for field of view constraints was defined based on the feature point coordinatesin the image space. The potential fields for target deflection and the relative pose were calculated based onthe scaled three-dimensional reconstruction results obtained from the decomposition of homography. Therobot can be regulated to the desired pose with the targets kept in the field of view under the hybrid artificialpotential field. Simulation results were provided in some representative circumstances to show theeffectiveness of the proposed method.