A 3D Reactive Collision Avoidance Algorithm for Nonholonomic Vehicles

Abstract

This paper presents a 3D reactive collision avoidance algorithm for vehicles with nonholonomic constraints. The algorithm steers the heading and pitch angle of the vehicle in order to maintain a constant avoidance angle to the obstacle, thus ensuring a safe collision avoidance maneuver. The flexibility provided by moving in three dimensions is utilized by choosing an optimal pair of safe pitch and heading angles for avoidance. Furthermore, the algorithm incorporates limits on the allowed pitch angle, which are often present in practical scenarios. The collision avoidance property is mathematically proved, and the analysis is validated by several numerical simulations.

Deep Reinforcement Learning Controller for 3D Path Following and Collision Avoidance by Autonomous Underwater Vehicles

“…In this sense, the reactive methods are considered local and are used when rapid action is required. Examples of reactive methods are the dynamic window approach ( Fox et al, 1997 ; Eriksen et al, 2016 ), artificial potential fields ( Williams et al, 1990 ), and constant avoidance angle ( Wiig et al, 2018 ). A potential pitfall with reactive methods is local minima manifested as dead ends ( Eriksen et al, 2016 ).…”

“…Additionally, extensive research discusses methods for COLAV in 2D that cannot be directly applied to 3D. In many cases where such methods are adapted to 3D, however, they do not optimally take advantage of the extra dimension ( Wiig et al, 2018 ).…”

A 3D reactive collision avoidance algorithm for underactuated underwater vehicles

“…The work builds on the preliminary contributions presented in Krogstad (2018a, 2018b). The 3D cone of safe directions keeping a constant avoidance angle to the obstacle is introduced in Wiig et al (2018a). Pitch limitations, which are often present for safety reasons in vehicles such as AUVs, are included in the algorithm, which operates on a kinematic vehicle modeled with nonholonomic constraints and limitations on the turning and pitching rates.…”

“…Only the vehicle kinematics was considered in Wiig et al (2018a). In Wiig et al (2018b), the design and analysis of the algorithm also included the important underactuated dynamics of an underactuated marine vehicle.…”

“…By smoothing the required pitch and yaw rate during the discrete switch from nominal operation when starting the collision avoidance maneuver, we ensure that all the control signals in the system remain well defined. Furthermore, while Wiig et al (2018aWiig et al ( , 2018b only considered static obstacles, we will in this paper extend the algorithm to also include moving obstacles by compensating the cone of safe directions for the obstacleʼs velocity. The compensation is dependent also on the vehicle speed, which contains an underactuated component when the vehicle is turning.…”