A new USC paper, Hands-off Robotics for Post-Stroke Arm Rehabilitation (PDF format), describes an autonomous assistant mobile robot that helps post-stroke patients during their recovery. The speed of recovery of such patients is dependent on the amount of exercise of the affected limb. The robot monitors the patient's limb using a variety of on-board sensors such as a camera and sonar as well as a wearable IMU. The robot monitors the patients activity and reminds or encourages the patient to use their arm as needed. The robot uses a behaviour-based software approach and is built on the standard Player/Stage software package. This research was done by Jon Eriksson. For more info see the project description or the project page at the USC Interaction Lab, which includes photos of the robot interacting with patients.