Adaptation for Validation of a Consolidated Control Barrier Function based Control Synthesis
Mitchell Black and Dimitra Panagou
Submitted to 2023 IEEE Conference on Robotics and Automation (under review)
We develop a novel adaptation-based technique for safe control design in the presence of multiple control barrier function (CBF) constraints. Specifically, we introduce an approach for synthesizing any number of candidate CBFs into one consolidated CBF candidate, and propose a parameter adaptation law for the weights of its constituents such that the controllable dynamics of the consolidated CBF are non-vanishing. We then prove that the use of our adaptation law serves to certify the consolidated CBF candidate as valid for a class of nonlinear, control-affine, multi-agent systems, which permits its use in a quadratic program based control law. We highlight the success of our approach in simulation on a multi-robot goal-reaching problem in a crowded warehouse environment, and further demonstrate its efficacy experimentally in the laboratory via AION ground rovers operating amongst other vehicles behaving both aggressively and conservatively.
In this video, a ground rover safely reaches its goal location using a consolidated control barrier function (C-CBF) based control synthesis in the presence of two non-responsive robots, one static and one dynamic. See the paper on arXiv (https://arxiv.org/abs/2209.08170) for more information.
This experiment was conducted in the Distributed Aerospace Systems and Control (DASC) Lab at the University of Michigan.
In a warehouse environment, 3 bicycle robots controlled by a decentralized consolidated CBF-QP control law (blue) navigate through a narrow corridor and busy intersection filled with non-responsive agents (red) en route to their goals.
A swarm of bicycle robots, each of whom use a decentralized consolidated CBF-QP control law, begin at randomly gridded locations and maneuver around each other to reach their goals while remaining inside the restricted area.
