Persistent Human Control in a Reservation-Based Autonomous Intersection Protocol

Karl Bentjen (1) , Scott Graham (1) , Scott Nykl (1)

1 Air Force Institute of Technology

Widespread use of fully autonomous vehicles is near. However, the desire of human beings to maintain control of their vehicles – even limited control – is unlikely to ever go away. Several protocols (e.g., AIM, SemiAIM and H-AIM) have been developed to safely and efficiently manage reservation-based intersections with a mixture of fully autonomous, semi-autonomous and non-autonomous vehicles. However, these protocols do not incorporate the dynamic of a human maintaining control of a semi-autonomous vehicle when approaching and crossing an intersection. This chapter lays the foundation for the extensions required for human-control of semi-autonomous vehicles, the ultimate goal being a protocol that maintains the efficiency of a fully autonomous environment while allowing human control of vehicles when navigating an intersection. This chapter also proposes information feedback mechanisms for human response, such as displays that provide the intersection arrival time, goal velocity, lane maintaining assistance and other warnings. Additionally, it describes a synthetic environment that enables the testing of intersection protocols that support human interaction.

Mots clés: Semi-autonomous vehicles; intersections; reservations; human control

Domaines: Informatique [cs]

Karl Bentjen, Scott Graham, Scott Nykl. Persistent Human Control in a Reservation-Based Autonomous Intersection Protocol. 12th International Conference on Critical Infrastructure Protection (ICCIP), Mar 2018, Arlington, VA, United States. pp.197-212, ⟨10.1007/978-3-030-04537-1_11⟩. ⟨hal-02076295⟩