Journal
IEEE ROBOTICS AND AUTOMATION LETTERS
Volume 8, Issue 9, Pages 6006-6012Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LRA.2023.3301297
Keywords
Robot safety; aerial systems; mechanics and control; model predictive path integral control; control barrier function; collision avoidance
Categories
Ask authors/readers for more resources
This letter presents a novel predictive control algorithm that combines control barrier function (CBF) technique with path integral control to develop a safety-critical controller for quadrotors. An efficient sequential projection algorithm is adopted to steer unsafe sampled inputs to a feasible region limited by safety barrier certificates. Comparisons with existing methods show that the proposed controller, which provides collective thrusts and body rates (CTBR) control policies, is able to deliver more robust maneuvers in obstacle avoidance.
The safety issue arises as one of the most important requirements for autonomous quadrotor flight. Recently, control barrier function (CBF) technique has been developed to provide necessary and sufficient conditions on safety for robot systems. Previous works unified CBF with control techniques in an optimization fashion. However, these methods could not provide consistent performance, especially in clutter environments. In this letter, we present a novel predictive control algorithm that combines CBF and path integral control to develop a safety-critical controller for quadrotors. An efficient sequential projection algorithm is adopted to steer unsafe sampled inputs to a feasible region limited by safety barrier certificates. Comparing to existing methods, one distinguishing feature of our proposed method is that we provide collective thrusts and body rates (CTBR) control policies. Simulations and experiments demonstrate that our controller can deliver more robust maneuvers in the course of obstacle avoidance than state-of-the-art safe controllers.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available