Robot Task Planning Using First-Order Logic
Abstract Task planning for a robot operating in an unknown environment using first-order logic is considered in this study. The approach is to use one agent to simulate the robot and a second agent to simulate the environment. Both agents employ knowledge bases and an inference engine. The rules for the knowledge bases are developed using first-order logic and the inference method is based on hyper-resolution. A weighting scheme is used by the robot to decide on the action to be taken. After enough domain information is obtained, a task planner, which is also based on rules, is employed. Simulation results validating the methodology are presented for a robot moving inside a warehouse with four rooms. In this example, the environment is initially unknown to the robot. But after mapping the environment, the robot can efficiently plan tasks such as moving an object from one room to another.