安防巡逻机器人全覆盖视觉导航方法

Visual Navigation Method for Security Patrol Robot with Full Coverage

为实现安防巡逻机器人视觉区域高效覆盖,准确跟踪巡逻路径,提出了基于多层代价地图的机器人全覆盖视觉导航方法。通过图像预处理、检测障碍物边缘、提取障碍物特征点三个步骤构建机器人安防巡逻环境模型,将周边区域中活动行为看作人造引力场运动,运用逆势场法优化人工势场法的局部最优解问题,通过直线连接原始方位与全部切线点,计算机器人运行的安全路径;行人感知前提下组建个人空间与群组交互社会代价模型,运用A*算法在导航搜索图内实施启发式搜索,并根据路径长短、社会收敛代价计算真实社会代价,增强导航准确性。测试结果表明,所提方法可提高机器人多变环境适应能力,导航实时性强,精度较高。

In order to achieve efficient coverage of visual area and accurate tracking of patrol path of security patrol robot,a robot full coverage visual navigation method based on multi-layer cost map is proposed. The robot security patrol environment model is constructed through three steps: image preprocessing,detecting the edge of obstacles and extracting the characteristic points of obstacles. The activity in the surrounding area is regarded as the movement of artificial gravitational field. The local optimal solution problem of artificial potential field method is optimized by using the inverse potential field method,and the safe path of robot operation is calculated by connecting the original orientation and all tangent points through a straight line;Under the premise of pedestrian perception,the social cost model of personal space and group interaction is established,the A * algorithm is used to implement heuristic search in the navigation search map,and the real social cost is calculated according to the path length and social convergence cost to enhance the navigation accuracy. The test results show that the proposed method can improve the adaptability of the robot to changeable environments,has strong real-time navigation and high precision.