Improving Oscillation & Navigation in Complex Environments Based on ROS2 Through Path Modification

In autonomous driving, path planning is a critical technology that enables agents to reach their destinations without human intervention. To date, various algorithms have been proposed for path planning, including $\mathrm{A}^{*}$, Dijkstra, and RRT. However, these algorithms primarily focus on planning the shortest path, which can lead to issues in obstacle-dense environments. In such cases, the planned path may either fail to navigate obstacles and come to a halt or exhibit oscillation phenomena as the agent repeatedly attempts to bypass obstacles. This oscillation significantly degrades driving performance in autonomous systems. Therefore, this paper proposes a method to detect oscillation-prone areas using the planned path and cost maps and subsequently modify the path in those regions to ensure safe passage through such zones. Ultimately, the modified path demonstrated more stable performance, unaffected by obstacles, compared to conventional path planning methods.