A robot decision making framework using constraint programming

An intelligent robotic system must be capable of making the best decision at any given moment. The criteria for which task is “best” can be derived by performance metrics as well as the ability for it to satisfy all constraints upon the robot and its mission. Constraints may exist based on safety, reliability, accuracy, etc. This paper presents a decision framework capable of assisting a robotic system to select a task that satisfies all constraints as well as is optimized based upon one or more performance criteria. The framework models this decision process as a constraint satisfaction problem using techniques and algorithms from constraint programming and constraint optimization in order to provide a solution in real-time. This paper presents this framework and initial results provided through two demonstrations. The first utilizes simulation to provide an initial proof of concept, and the second, a security robot demonstration, is performed using a physical robot.