Towards Seamless Human-AV Interactions: Exploring Novel Paradigms for Teleoperation of Autonomous Vehicles


A Workshop at AutoUI'24 - September 22nd, 2024 - Stanford, CA, USA

Building a Framework for Teleoperation

Autonomous vehicle (AV) technologies aim for fully self-driving cars but still require human intervention for edge-case road scenarios. Remote human operation can mitigate AV shortcomings through two major teleoperation paradigms: remote driving (tele-driving) and remote assistance (tele-assistance). Tele-driving involves a remote operator (RO) continuously controlling the vehicle, while tele-assistance allows the RO to make high-level decisions by delegating the execution to automation. Tele-driving is complex, mostly due to the physical disconnect between the RO and the AV and latency. Tele-assistance faces unresolved design, implementation, and regulatory challenges. Examples for application are depicted below. This workshop provides a platform for researchers and practitioners to explore teleoperation challenges and opportunities. Participants will address teleoperation methods and design innovative remote AV operation paradigms through interactive activities, discussions, and poster sessions. The workshop aims to create a teleoperation paradigm-scenario mapping and generate novel concepts and interaction methods for resolving edge-case road scenarios.

Potential teleoperation applications can involve various actors on different levels. This allows for interaction concepts beyond traditional manual vehicle control (tele-driving with steering wheel and pedals, left) - for example, control via higher-level commands (tele-assistance, right).