Internal State-Based Risk Assessment for Robots in Hazardous Environment

Robots operating in a hazardous environment should be able to assess the risks involved before executing any given task. Though several risks can be considered in such a scenario, the risks based on environmental conditions are rarely explored. In this paper, we present a novel, risk-based navigation approach for robots in hazardous environments. We specifically investigate the environmental risks which can be measured and whose maximum state can be defined. These risks are integrated into the costmap of the robot to alter its traversability cost based on the robot’s current state. In doing so, the robot can adjust its path to account for hazards it has encountered on the fly. We term this approach as the Internal State-based Risk Assessment framework. We validate this framework using simulations where a robot must navigate through a nuclear environment whilst optimizing its path to avoid high radiation and high-temperature zones. We show that the robot can alter its path to account for the encountered hazards that have been mapped onto its internal state.