In context
In early 2023, outdoor mobile robots increasingly relied on GNSS-based path collection for autonomous navigation and map building. However, manually driven paths often contained invalid segments—such as zigzag patterns from repeated reversing—that rendered them unusable for trajectory following. Existing smoothing methods could not automatically identify and remove these invalid parts.
What was reported
Researchers from Army Military Transportation University in Tianjin proposed a two-stage method for processing GNSS paths collected by wheeled mobile robots. First, a DRD (drive-reverse-drive) segmentation algorithm divides the path based on heading changes, then eliminates invalid segments using rules for reverse-only and repeated-reverse patterns. Remaining segments are adaptively re-split and reconnected to form a clean path.
Second, a multi-objective optimization is formulated with path smoothness and fidelity to the original path as objectives, subject to rectangular region constraints and asymptotic endpoint constraints to ensure curvature continuity. The problem is transformed into a quadratic programming form for efficient solving.
Experiments on a real unmanned vehicle with RTK-GNSS showed the method effectively handled paths with one to three reversing events across straight, right-angle, and curved roads. The average position deviation between the optimized and rearranged path was less than 0.2 m, with an average processing time of 8.8 ms.
Why it mattered
This work addressed a practical bottleneck in automating map collection and trajectory generation for outdoor robots. By automatically cleaning invalid GNSS paths and producing smooth, drivable trajectories in near real time, it reduced manual intervention and improved the reliability of autonomous navigation in complex environments—an important step toward fully automated path acquisition for industrial and military applications.
The processed path can be applied to trajectory following of unmanned vehicle.
Source: 《机器人》期刊 (robot.sia.cn) · Published 2023-03-06 · “基于重组优化的轮式移动机器人路径处理方法”