A Genetic Algorithm Based Adaptive and Hybrid Pathfinding Algorithm to Enhance the Performance of Robotic Bin-Picking

Efficient robot path planning in obstacle-strewn environments remains a challenging task, especially in bin-picking applications. In the industrial context, bin picking involves the high-speed, accurate, and autonomous handling of randomly placed objects from bins. Enhancing path planning strategies in this area can significantly improve production efficiency, reduce direct labor costs, and improve production rates in automated processes. In this paper, we present a Genetic Algorithm based Adaptive Hybrid Pathfinding Algorithm (AHPFA) that computes the shortest, collision-free path for an industrial robot, thereby minimizing the total time required for path execution and improving overall operational efficiency. The algorithm was tested in simulated and real environments to evaluate its efficacy in finding the shortest path and ensuring collision-free trajectories in 3D obstacle-dense settings. The results from our real-world robot experiments highlight the superior performance of our proposed GA-based AHPFA in terms of the time required for path execution as compared with some of the other metaheuristic algorithms like PSO, ACO, and GA in navigating complex environments in real-time, with each algorithm found to offer distinct advantages and tradeoffs. The proposed algorithm's performance was also assessed based on convergence rate, computational efficiency, and path length metrics. This work contributes to developing more robust and efficient path-planning frameworks, enhancing the ability of robotic systems to operate effectively in complex industrial environments.