Multi-sensor fusion architecture for human-autonomy teaming

The adaptation of a sensor fusion architecture typically utilized in multi-sensor autonomous systems toward a human-autonomy driving system is presented. The adapted sensor fusion architecture employs a simplified methodology for estimating the expected and actual performance of human and autonomous counterparts in a joint task. The aim of this estimation strategy is to mitigate human bias and miscalibrated trust-in-autonomy (TiA), through real-time visual feedback to the human subject to recommend authority toggling between the agents, and thereby increase system resilience.