We present an approach for autonomous sensor control for information gathering under partially observable, dynamic and sparsely sampled environments. We consider the problem of controlling a sensor that makes partial observations in some space of interest such that it maximizes information about entities present in that space.
We present ROSRV, a runtime verification framework for robotic applications on top of the Robot Operating System (ROS), a widely used open-source framework for robot software development. ROSRV aims to address the safety and security issues of robots by providing a transparent monitoring infrastructure that intercepts and monitors the commands and messages passing through the system. Safety and security properties can be defined in a formal specification language, and are ensured by automatically generated monitors. ROSRV integrates seamlessly with ROS—no change in ROS nor the application code is needed. ROSRV has been applied and evaluated on a commercial robot.