Protective Autonomy For Improved Remote Control Of Robot-vehicles

Intelligent robots will possess data processing activities which will enable them to respond to information with behaviour which appears to be optimal with respect to preset goals. The information will include both state and process. The robot’s intelligence will be a function of the amount, variety, and complexity of its informational input, its storage capacity, the number and complexity of its goals, the degree of optimality achieved in attaining these goals, the amount and complexity of the output instructions involved in the response, and the position on a scale strategic/tactical of the activity involved. Intelligent robots in the sense outlined above do not yet exist, and when they do they will need to possess many autonomous skills. To meet this challenge an autonomous skill which will benefit robot-vehicles is one which gives the system the ability to survive. This skill is summarised by the term Protective Autonomy. The long term goal of this programme of work is to provide a robot-vehicle with algorithms which support the strategic behaviour of homing to a spatial goal that is beyond the sensor horizon, and which concurrently utilise tactical observations made from within the sensor horizon to ensure that the execution of the output instructions result in both the safe movement within the sensor horizon and the achievement of the strategic goal seeking