Project E2 designs safe human-to-machine handover techniques in mixed-initiative control and addresses the recent trend to human oversight in autonomous vehicles and aviation. If a critical situation arises, operators need to be supported by an efficient interface for regaining and maintaining situational awareness during a handover. The research goals are 1) to extend the simulation environment to a physical lab representing a control centre for human oversight of multiple drones, 2) in the case of a critical situation, to explain not just the current system state, but also relevant information from the past and the foreseeable future, and 3) to develop suitable multimodal output techniques to explain the situation to the operator, also considering multiple drones.
Previous Work (2019-2022)
In mixed-initiative control, one important issue is handover. The mission of this project is to better understand how machines can safely hand over control to humans, and to develop technological support. We will use formal methods from AI – description logic (DL) and automated planning – to more reliably predict when a handover is necessary, and to increase the advance notice for handovers by planning ahead at run-time. We will combine methods from human-computer interaction and natural language generation to develop solutions for safe and smooth handovers. Ultimately, we strive to tightly integrate HCI and formal methods, making human aspects of the human-machine system more accessible to formal analysis, thereby ensuring operational safety.