A Shared Autonomy Approach for Wheelchair Navigation Based on Learned User Preferences

Research on robotic wheelchairs covers a broad range from complete autonomy to shared autonomy to manual navigation by a joystick or other means. Shared autonomy is valuable because it allows the user and the robot to complement each other, to correct each other's mistakes and to avoid collisions. In this paper, we present an approach that can learn to replicate path selection according to the wheelchair user's individual, often subjective, criteria in order to reduce the number of times the user has to intervene during automatic navigation. This is achieved by learning to rank paths using a support vector machine trained on selections made by the user in a simulator. If the classifier's confidence in the top ranked path is high, it is executed without requesting confirmation from the user. Otherwise, the choice is deferred to the user. Simulations and laboratory experiments using two path generation strategies demonstrate the effectiveness of our approach.