An algorithm will be presented that ensures safe navigation in an environment in which other vehicles are present. Each vehicle knows its own position and speed and the position and speed of the other vehicles within a certain distance. Each vehicle can assume that the other vehicles try to avoid collision in the same way as own vehicle. It does not know the destination of other vehicles. The space in which a collision manoeuvre takes place is classified in five categories. For each category a collision avoidance strategy is specified, which determines how vehicles pass each other and gives a control law which specifies the actual speed and course alterations during the manoeuvre. Special attention is paid to the problem of navigating in a corridor and how to decide whether to overtake a slower vehicle ahead or not.

Algorithms for the Avoidance of ColIision between Autonomous Mobile Robots

Distributed Control for Collision Avoidance between Autonomous Vehicles

https://ieeexplore.ieee.org/iel2/4535/12853/00593271.pdf

In the last number of the Journal (14, 243) Hollingdale showed mathematically that anticollision manoeuvres for the case of two craft moving in a plane can be formulated on a rigorous logical basis; and that a simple set of manoeuvres could be devised to meet all the geometrical requirements for anti-collision. The convention adopted was that each craft should manoeuvre so that if the other craft stood-on, the sight-line (the line joining the two craft) would always rotate in an anti-clockwise direction. The analysis showed that, mathematically, this was the complete and only answer to the basic problem. The author of the present paper compares systems, such as that represented by the Collision Regulations and the Rules of the Air, based on onus to avoid, with a system based on a convention as to the direction of rotation of the sight-line. The first type of system, he holds, is defective for visual sightings and almost useless for radar sightings. The system put forward in the paper can be applied to both visual and radar sightings whether unilateral or bilateral by orientating a simple diagram with own course. A means of inter-communication is desirable but even without it such a system can, the author believes, safely be employed in the greater proportion of encounters. The collision problem as a whole is beset by operational factors which are entirely distinct from the geometry of the situation, and for that reason no purely mathematical approach can, in isolation, be entirely valid. Nevertheless, there can be no doubt that studies such as the present one, based on a proper understanding of the geometry, can contribute greatly to the solution of this most pressing problem of navigation.

Algorithms for the Avoidance of ColIision between Autonomous Mobile Robots

Citations

Distributed Control for Collision Avoidance between Autonomous Vehicles

References

A Comparison of Two Systems for Avoiding Collision

Related Papers (5)

A navigation of mobile robots with collision avoidance for moving obstacles

Obstacle avoidance for mobile robots based on relative coordinates

Cooperative Navigation Function based Navigation of Multiple Mobile Robots

Trajectory Planning and Collision Avoidance Algorithm for Mobile Robotics System

Sensor management for local obstacle detection in mobile robots