Single-Target Point Navigation for Indoor Differential-Drive Vehicles: Integrating Constrained-Sampling Rapidly Exploring Random Tree and Dynamic Window Approach Algorithm

Path planning for indoor differential-drive vehicles in cluttered, dynamic environments is challenging, primarily due to the inherent trade-off between global path optimality and local obstacle avoidance responsiveness. To address this, we propose an integrated navigation framework that combines a constrained-sampling Rapidly Exploring Random Tree (RRT*) algorithm for efficient global path generation with an enhanced Dynamic Window Approach (DWA) incorporating a global path consistency term into its trajectory evaluation function. The method is validated through extensive simulations and physical experiments on a WHEELTEC two-wheel differential-drive robot. The results demonstrate that the proposed approach achieves a 100% task success rate across diverse scenarios; significantly reduces both path length and travel time compared with baseline methods; and effectively prevents stagnation in complex obstacle configurations, such as U-shaped traps, by guiding local motion toward the global path. These improvements highlight the benefit of tightly coupling global guidance with local planning. The framework provides a robust and efficient solution for autonomous navigation in indoor differential-drive vehicles.