Anticipatory stigmergic collision avoidance under noise

Abstract

Reactive path planning to avoid collisions with moving obstacles enables more robust agent systems. However, many solutions assume that moving objects are passive; that is, they do not consider that the moving objects are themselves re-planning to avoid collisions, and thus may change their trajectory. In this paper we present a model, Anticipatory Stigmergic Collision Avoidance (ASCA) for reciprocal collision avoidance using anticipatory stigmergy. Unlike standard stigmergy, in which agents leave pheromones to indicate a trace of previous actions, anticipatory stigmergy deposits pheromones on intended future paths. By sharing their intended future paths with each other at regular intervals, agents can re-plan to attempt to avoid collisions. We experimentally evaluate ASCA over three scenarios, and compare with a state of art approach, Reciprocal Velocity Obstacles (RVO). Our evaluation showed that ASCA is consistently more robust in noisy environments in which transmitted information can be lost or degraded. Further, using ASCA without noise results in fewer collisions than RVO when agents are in formation, but more collisions when formed randomly.

Multi robot collision avoidance in a shared workspace

“…Other approaches for multi-robot collision avoidance use auctions (Calliess et al 2012) at a rather high communication overhead, or stigmergy (Theraulaz and Bonabeau 1999;Lemmens and Tuyls 2012;von der Osten et al 2014), which relies on pheromones that are hard to apply in a real world setting. Additionally, these approaches do not implement robot-human avoidance.…”

Stigmergy-based collision-avoidance algorithm for self-organising swarms

“…The optimised escape trajectory approach consists of solving an optimisation problem that combines the drone kinematic model with a set of physical and behavioural constraints. The following decentralised multi-agent algorithms fall within this group: 1) the Reciprocal Velocity Obstacles (RVO) and the derived Optimal Reciprocal Collision-Avoidance (ORCA) algorithms [2]; 2) the Cooperative Dynamic (CoDy) algorithm [15], which is able to solve dead-lock situations; 3) the Context-Aware Route Planning (CARP) algorithm [4,13], which is a graph routing algorithm aimed at finding the shortest trajectory and avoiding collisions; and 4) other optimisation-based algorithms such as [14] and [20].…”

“…Others try to infer the state of the overall swarm through stigmergy, assuming this would be useful for collision avoidance purposes. For example, an approach based on anticipatory stigmergic collision avoidance (ASCA) under noise is proposed in [14], which consists of using pheromone information in a rather unusual fashion: instead of leaving a trail of pheromones over past positions, all drones in the swarm share information in the same indirect manner but about future intended positions instead. The drones will then optimise their trajectories in order to avoid locations with high concentrations of pheromone.…”

“…Numerous CASs have been proposed in the literature, the majority of which are suitable for Autonomous Ground Vehicles (AGVs) only. This is because CASs typically rely on two main assumptions [14]: 1) stationary vehicles (i.e. vehicles can stop suddenly and remain still indefinitely) and 2) perfect information (accurate noiseless perception).…”