Risk-based sensing in support of adjustable autonomy

Abstract

Current unmanned systems are typically tele-operated and manpower intensive relying on human operators and their decision-making capabilities to perform platform and mission tasks. We envision a future where unmanned platforms have greater decision making abilities and autonomous behaviors are the norm. For example, unmanned autonomous systems will be deployed as teams along with their human supervisor. The UAS will interact with the supervisor at a high level, then take on an expanded role in mission planning, resource allocation and route planning. Toward this end, we have developed a risk-based adjustable autonomy system with a task directed adaptive sensing technology concept to allow system autonomy operation at a level in which an operator has confidence of success. We test our idea in simulation for a natural disaster recovery task. We present experimental results verifying the utility of our technology.