Autonomous risk-aware exploration

This paper considers the problem of an autonomous robot being switched on and making its first movements in a totally unknown environment. The robot has to be cautious because due to the lack of information about the environment. Any movement might be a potential risk, e.g. colliding with an obstacle. We present a novel approach to determine the next best movement of an autonomous robot based on the current knowledge about the enclosing world. The problem of exploring an unknown environment is done by maximizing the knowledge of the environment and minimizing the risk of colliding, taking in account the uncertainty of the sensor measurements and the so far obtained knowledge.