Dr. Stefano V. Albrecht awarded RAEng Industrial Fellowship
Information about Dr. Albrecht's fellowship research
Multi-agent reinforcement learning for warehouse logistics with robotic and human co-workers
We envision a futuristic warehouse in which dozens or hundreds of mobile robots and humans work together to collect and deliver ordered items inside a large-scale warehouse. Existing approaches for order-picking systems in industry based on fixed heuristics require a large engineering effort to operate under specific warehouse layouts and resource constraints, and their achievable performance is often limited by heuristic design limitations. To improve the efficiency and flexibility of warehouse order-picking systems, we will leverage research in Multi-Agent Reinforcement Learning (MARL) in which the workers learn directly through trial-and-error how to optimally collaborate with one another in the warehouse. This project will develop and demonstrate MARL-based solutions in real-world warehouse systems involving robotic and human workers through a close partnership with Dematic/KION, a global leader in warehouse automation technologies.