Visual Model Building for Robot Sensemaking: Perspectives, Challenges, and Opportunities

Chiatti, Agnese; Bardaro, Gianluca; Matteucci, Matteo and Motta, Enrico (2023). Visual Model Building for Robot Sensemaking: Perspectives, Challenges, and Opportunities. In: Bridge Session on AI and Robotics of the thirty-seventh AAAI conference on Artificial Intelligence (AAAI-23), 7-8 Feb 2023, Washington DC, USA, AAAI.

URL: https://aaai.org/Conferences/AAAI-23/

Abstract

Robots can help with many visually-intense and onerous tasks that are traditionally carried out by human workers, such as the inspection of critical infrastructures or the management of crops. However, before we can safely delegate tasks to robots, we ought to ensure that they can reliably make sense of the environments in which they are deployed. To this aim, building world models that resemble the complexity of the real world is critical. Despite research efforts in the AI and Robotics communities towards tackling the problem of model building, few works exist that approach this problem by considering perspectives and lessons learned from both fields. In this position paper, we use three strategic application domains in Robotics to argue for the centrality of visual model building to support robot sensemaking.

Viewing alternatives

Item Actions

Export

You can export this page using these formats