Integrating planning and control for single-bodied wheeled mobile robots

Abstract

This paper presents an approach to couple path planning and control for mobile robot navigation in a hybrid control framework. We build upon an existing hybrid control approach called sequential composition, in which a set of feedback control policies are prescribed on well-defined domains contained in the robot's free space. Each control policy drives the robot to a goal set, which lies in the domain of a subsequent policy. Control policies are deployed into the free state space so that when composed among one another, the overall action of the set of control policies drives the robot to perform a task, such as moving from a start to a goal location or patrolling a perimeter. A planner determines the sequence of control policies to be invoked. When control policies defined in this framework respect the low-level dynamics and kinematics of the system, this formal approach guarantees that high-level tasks are either accomplished by a given set of policies, or verifies that the tasks are not achievable with the given policies. Specifically, this paper makes three key contributions to the technique of sequential composition. First, we allow for a larger variety of transitions among the control policies, which results in a richer set of generated behaviors. In particular, we are able to design policies that respect multiple interacting constraints including obstacles, nonholonomic constraints, and input bounds. This is achieved by utilizing flow-through control policies, which allow the robot to exit the policy domain through a subset of the domain boundary. Second, our hybrid control approach lifts the planning problem from working in trajectory space to control policy space, allowing for both more efficient and meaningful symbolic planners to be employed for systems with nonholonomic constraints. Finally, we address the issue of deploying control policies in configuration spaces whose obstacles have curvature (i.e., are not polytopes in configuration space); ultimately, such an approach results in the collection of policy domains only approximating the free space. We believe this trade-off is necessary, and provide a technique to measure the relative completeness of the approximation. The paper concludes with an experimental validation of our approach on a mobile robot.