Title

Local Pursuit Strategy-Inspired Cooperative Trajectory Planning Algorithm For A Class Of Nonlinear Constrained Dynamical Systems

Keywords

Bio-inspired control; cooperative control; nonlinear constrained optimisation

Abstract

Cooperative trajectory planning is crucial for networked vehicles to respond rapidly in cluttered environments and has a significant impact on many applications such as air traffic or border security monitoring and assessment. One of the challenges in cooperative planning is to find a computationally efficient algorithm that can accommodate both the complexity of the environment and real hardware and configuration constraints of vehicles in the formation. Inspired by a local pursuit strategy observed in foraging ants, feasible and optimal trajectory planning algorithms are proposed in this paper for a class of nonlinear constrained cooperative vehicles in environments with densely populated obstacles. In an iterative hierarchical approach, the local behaviours, such as the formation stability, obstacle avoidance, and individual vehicle's constraints, are considered in each vehicle's (i.e. follower's) decentralised optimisation. The cooperative-level behaviours, such as the inter-vehicle collision avoidance, are considered in the virtual leader's centralised optimisation. Early termination conditions are derived to reduce the computational cost by not wasting time in the local-level optimisation if the virtual leader trajectory does not satisfy those conditions. The expected advantages of the proposed algorithms are (1) the formation can be globally asymptotically maintained in a decentralised manner; (2) each vehicle decides its local trajectory using only the virtual leader and its own information; (3) the formation convergence speed is controlled by one single parameter, which makes it attractive for many practical applications; (4) nonlinear dynamics and many realistic constraints, such as the speed limitation and obstacle avoidance, can be easily considered; (5) inter-vehicle collision avoidance can be guaranteed in both the formation transient stage and the formation steady stage; and (6) the computational cost in finding both the feasible and optimal solutions is low. In particular, the feasible solution can be computed in a very quick fashion. The minimum energy trajectory planning for a group of robots in an obstacle-laden environment is simulated to showcase the advantages of the proposed algorithms. © 2013 Taylor & Francis.

Publication Date

3-4-2014

Publication Title

International Journal of Control

Volume

87

Issue

3

Number of Pages

506-523

Document Type

Article

Personal Identifier

scopus

DOI Link

https://doi.org/10.1080/00207179.2013.845911

Socpus ID

84893941582 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/84893941582

This document is currently not available here.

Share

COinS