Automating Geometric Proofs of Collision Avoidance with Active Corners

Avoiding collisions between obstacles and vehicles such as cars, robots, or aircraft is essential to the development of autonomy. To simplify the problem, many collision avoidance algorithms and proofs consider vehicles to be a point mass, though the actual vehicles are not points. In this paper, we...

Full description

Saved in:
Bibliographic Details
Published in2022 Formal Methods in Computer-Aided Design (FMCAD) pp. 1 - 10
Main Authors Kheterpal, Nishant, Tang, Elanor, Jeannin, Jean-Baptiste
Format Conference Proceeding
LanguageEnglish
Published FMCAD Association and authors 18.10.2022
Subjects
Aircraft
Automobiles
Benchmark testing
Collision avoidance
Design automation
Robots
Trajectory
Online AccessGet full text

Cover

More Information
Summary:Avoiding collisions between obstacles and vehicles such as cars, robots, or aircraft is essential to the development of autonomy. To simplify the problem, many collision avoidance algorithms and proofs consider vehicles to be a point mass, though the actual vehicles are not points. In this paper, we consider a convex polygonal vehicle with nonzero area traveling along a 2-dimensional trajectory. We derive an easily-checkable, quantifier-free formula to check whether a given obstacle will collide with the vehicle moving on the planned trajectory. We apply our active corner method to two case studies of aircraft collision avoidance and benchmark its performance.
ISSN:2708-7824
DOI:10.34727/2022/isbn.978-3-85448-053-2_43