A convex-lifting approach to reactive path planning

This paper presents a reactive path planning framework based on convex lifting for safe navigation in dynamic environments. The framework employs an iterative selective path planning strategy, in which a new candidate path is generated at each time step and updated (guaranteeing improvements by direct or indirect comparison with the previous one). A novel method, termed set-interpolation , is then introduced to adaptively enlarge obstacles within the convex-lifting generated partition. This mechanism allows flexible evolution of paths farther from the agent while conserving the path near the agent. The enlargement process effectively creates a dynamic buffer zone that shifts the agent’s safe partition away from obstacle boundaries, thereby enhancing safety in real-time operation. The proposed framework is evaluated in a time-varying environment with multiple moving obstacles.