复杂环境下机器人多源协同递推定位算法研究

Research on multi-source collaborative recursive localization algorithm for robot in complex environment

轮式机器人通常采用GNSS/INS组合导航来满足室外定位需求,但在复杂环境中,GNSS定位精度逐渐下降,且惯性导航模块的误差持续累积,导致机器人无法通过GNSS/INS组合导航系统获得精确的定位。为解决这一问题,从轮式机器人行驶环境的特征出发,设计了一种基于相机拍摄的车道线边缘检测算法,利用扩展卡尔曼滤波算法结合车道线信息,对车辆定位误差进行约束,并提出了一种适用于城市环境的GNSS/INS/相机多源融合定位算法。

Wheeled robots typically use GNSS/INS integrated navigation to meet outdoor positioning needs,but in complex environments,GNSS positioning accuracy gradually decreases,and errors in inertial navigation modules continue to accumulate,making it impossible for robots to obtain accurate positioning through GNSS/INS integrated navigation systems.To solve this problem,a camera based lane edge detection algorithm was designed based on the characteristics of the driving environment of wheeled robots.The extended Kalman filter algorithm combined with lane information was used to constrain the vehicle positioning error,and a GNSS/INS/camera multisource fusion positioning algorithm suitable for urban environments was proposed.