Enhancing Efficiency of Automated Medication Delivery via Reinforcement Learning in Hospital Settings

With the rapid advances in robotics and automation science, robots have been taking over more and more jobs that used to be performed by humans. Automated medication delivery with robots can free up nurses from the workload. Path planning for robots is a hot research topic, which is particularly important in hospital settings. In this paper, we propose to use a reinforcement learning approach to find optimal medication delivery paths. The optimization objective is to minimize the time span for delivering medications to all patient rooms. Practical factors such as collision avoidance are considered. To reduce the action space of the reinforcement learning agent, unreasonable actions are filtered to save the training time.