Despite the great progress in the field of robotic navigation over the past few decades, navigating a human environment remains a hard task for a robot, due to the lack of formal rules guiding traffic, the lack of explicit communication among agents and the unpredictability of human behavior. Existing approaches often result in robot motion that is hard to read, which causes unpredictable human reactions to which the robot in turn reacts to, contributing to an oscillatory joint behavior that hinders humans’ paths. We argue that the root of the problem lies in the failure from the robot’s part to convey consistently its intentions to human observers.
This project aims at developing an autonomous robotic system, capable of navigating crowded environments in a socially competent fashion. To this end, we develop novel models, algorithms, software and systems, which we plan on validating experimentally in real-world scenarios.
The main novelty of our approach lies in the development of a novel planning framework for closed-loop navigation in dynamic multi-agent workspaces. The core of the framework is a novel topological representation, based on braid groups, that models the collective behavior of multiple agents. Based on this representation and employing data-driven techniques, our algorithms generate motion plans that are consistent with the perceived context, thus resulting in socially competent robot behaviors that allow for smooth integration of mobile robots in crowded human environments. Our framework is inspired by insights from studies on pedestrian behavior and action interpretation and leverages the power of implicit communication to overcome the complications of the uncertainties induced by the imperfections of existing models of human decision making.
Our robot platform is a Suitable Technologies Beam Pro, equipped with on-board computation and an Occam Omni Stereo 360 RGBD camera, interfaced through ROS.