As a part of course assignment, the Social Forces model has been implemented for obstacle avoidance. The model is then tested in various bottleneck test cases.
The Algorithm :
The Algorithm :
The Social Forces model computes various forces exerted on a agent by the environment. It then adds the exerted force with the force required by the agent to reach the goal(Fpref) .
Fagents & Fwalls includes the proximity force, repulsion force(in case of collision) and sliding friction force(in case of collision) by obstacles/other agents.
Code : https://github.com/CG-F15-10-Rutgers/SteerLite (Forked from SteerLite parent repo)
The model is run against several test cases and the results are recorded. Following are the video links :
Futher improvements : The efficiency can be further improved by tweaking the constant parameters of the Social Forces algorithm as per the use-case.
F = Fgoal + Fagents + Fwalls
Fagents & Fwalls includes the proximity force, repulsion force(in case of collision) and sliding friction force(in case of collision) by obstacles/other agents.
Code : https://github.com/CG-F15-10-Rutgers/SteerLite (Forked from SteerLite parent repo)
The model is run against several test cases and the results are recorded. Following are the video links :
(a) Hallway one-way
(b) Hallway two-way
(c) Hallway four-way
(d) Bottleneck evacuation
No comments:
Post a Comment