Optimal Path Generation and Real-time Scheduling for Autonomous Mobile Platforms

Optimizing path planning and real-time scheduling are crucial for autonomous mobile vehicles under dynamic conditions. In this study, advanced algorithms are developed to achieve the above functions under breakdown scenarios for autonomous sweeping trucks with operational constraints. The methodology involves Spectral clustering, utilizing its ability to efficiently assign service routes to a predefined number of vehicles. Simulated Annealing (SA) in conjunction with the Traveling Salesman Problem (TSP) is then employed to systematically optimize route sequences, ensuring minimal travel distances and efficient coverage. Real-time scheduling functions dynamically redistribute routes using a comprehensive approach tailored to breakdown disruptions, ensuring operational efficiency. Edge redistribution employs Kernighan-Lin bisection to equally allocate remaining unattended road edges from broken vehicles to operational ones, promoting an equitable distribution of tasks among available resources. Simultaneously, Dijkstra's algorithm is applied to identify the shortest path between the last edge of a working vehicle and the first redistributed one, minimizing travel distances and optimizing route sequences. Reconnection strategies are implemented to eliminate any disconnected edges resulting from the redistribution process. Simulations demonstrate the developed algorithms' effectiveness, providing a 71.6% efficiency rate and a short computation time of 35.63 seconds for a 17.68 km service route. Additionally, detailed statistical calculations include normal service distance, deadhead distance, and overall efficiency for both normal operation and breakdown scenarios. The algorithms can enhance work efficiency and resource utilization for various autonomous mobile systems through adaptive path planning and real-time scheduling based on operational conditions and demand.