Fast and Efficient Motion Planning: using Hybrid Sampling and Simulated Annealing

This paper introduces a Modified Simulated Annealing (MSA) algorithm for mobile robot path planning that achieves an optimal balance between computational complexity and robustness. MSA occupies the ideal performance region with O ( n 1.2 ) Complexity, 95% success rate, and 82% sampling efficiency, significantly outperforming existing methods across diverse environmental complexities. Various path planning algorithms are currently employed to determine feasible paths for mobile robots in obstacle-filled environments. A review of the literature reveals that although many algorithms successfully generate collision-free paths, these consume a lot of samples to produce and result in increased path-finding time. Real-time path planning demands rapid convergence to a feasible solution by generating a minimum sample set. To address these limitations, this paper introduces a novel Modified Simulated Annealing (MSA) algorithm, which employs a systematic heapsort-based sampling approach and dynamically adjusts the randomness area based on temperature to escape local minima, ensuring faster convergence and higher efficiency, and to enhance global exploration. The algorithm is bench-marked against five sampling-based path-finding algorithms, evaluating success rate, computation time, path cost, and the number of samples generated on a consistent platform and workstation. The proposed approach demonstrated superior efficiency and reliability in terms of the number of samples and computation time compared to other methods. The generated path was simulated using the Unity6 platform, confirming collision-free motion and adherence to the planned trajectory. The paper further explores potential applications of the algorithm across aerospace, industrial, and commercial domains.