10581

Parallel multi-agent path planning in dynamic environments for real-time applications

Alexander Dooms
Department of Electronics and Information Systems, Faculty of Engineering and Architecture, University Ghent
University Ghent, 2013
@article{dooms2013parallel,

   title={Parallel multi-agent path planning in dynamic},

   author={Dooms, Alexander},

   year={2013}

}

Download Download (PDF)   View View   Source Source   

641

views

Current pathplanning algorithms are not efficient enough to provide optimal pathplanning in dynamic environments for a large number of agents in real time. Furthermore, there are no real-time algorithms that fully use the potential of parallelism. The goal of this thesis is to find a basis for such an algorithm. Based on the literature study, an algorithm is proposed which is based on a Quadtree roadmap. The Quadtree can rapidly be adapted to a dynamic environment, requiring only a couple of nodes to be re-evaluated. In addition, the pathplanning algorithm using the Quadtree takes full advantage of the fact that all nodes (and the zones of the environment they determine) are in the free space. Any path that stays within the zones is a valid path, and determining the optimal path is reduced to calculating with coordinates. A couple of optimizations have also been investigated, of which two stand out in particular. The principal shortcomings of the solution using Quadtrees is that we use a regular grid, with the links having a cost of 1 (or sqrt{2} for diagonal links), and that although each agent can search his path fully independently of the other agents, there is no parallelism during path search. Even with theses shortcomings, results show a great improvement in execution times compared to A*.
VN:F [1.9.22_1171]
Rating: 0.0/5 (0 votes cast)

* * *

* * *

TwitterAPIExchange Object
(
    [oauth_access_token:TwitterAPIExchange:private] => 301967669-yDz6MrfyJFFsH1DVvrw5Xb9phx2d0DSOFuLehBGh
    [oauth_access_token_secret:TwitterAPIExchange:private] => o29ji3VLVmB6jASMqY8G7QZDCrdFmoTvCDNNUlb7s
    [consumer_key:TwitterAPIExchange:private] => TdQb63pho0ak9VevwMWpEgXAE
    [consumer_secret:TwitterAPIExchange:private] => Uq4rWz7nUnH1y6ab6uQ9xMk0KLcDrmckneEMdlq6G5E0jlQCFx
    [postfields:TwitterAPIExchange:private] => 
    [getfield:TwitterAPIExchange:private] => ?cursor=-1&screen_name=hgpu&skip_status=true&include_user_entities=false
    [oauth:protected] => Array
        (
            [oauth_consumer_key] => TdQb63pho0ak9VevwMWpEgXAE
            [oauth_nonce] => 1474824349
            [oauth_signature_method] => HMAC-SHA1
            [oauth_token] => 301967669-yDz6MrfyJFFsH1DVvrw5Xb9phx2d0DSOFuLehBGh
            [oauth_timestamp] => 1474824349
            [oauth_version] => 1.0
            [cursor] => -1
            [screen_name] => hgpu
            [skip_status] => true
            [include_user_entities] => false
            [oauth_signature] => U42AP9aDVHYgyiFiK7+3L1lwtos=
        )

    [url] => https://api.twitter.com/1.1/users/show.json
)
Follow us on Facebook
Follow us on Twitter

HGPU group

1996 peoples are following HGPU @twitter

HGPU group © 2010-2016 hgpu.org

All rights belong to the respective authors

Contact us: