0
$\begingroup$

I am currently working on collision avoidance using Potential fields and I have visualized the same in winfom.

Now, I need to test the algorithm in predefined trajectory say a spline(or in a lane).

My question is: Is it possible to do collision avoidance considering potential field method in a predefined trajectory?

Say the model car avoids the obstacle and returns back on its trajectory to reach the goal, created in windows form. If, so what are the parameters and contraints should be considered?

Improve this question
$\endgroup$
1
$\begingroup$

This is not any more a pure potential field type of problem, since you already have a global trjectory which you then alter locally.

This is a hybrid approach. A good example for a hybrid approach is the Elastic Bands method, can be found here. It models the path as elastic bands which are deformed by virtual (mechanical) forces (due to the presence of obstacles). Furthermore, it also includes virtual forces which deform the path back to its original shape if no obstacles are present anymore, similarly to an elastic material.

Later this method has been alterd to include robot geometry also. The Elastic Strips Method can be found here.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Yes, I agree the approach has to be hybrid. I don't know about elastic bands, I have to check on that. But, my main concern is the ditance factor.Since, Potential field algorithm mainly uses distance to attract towards the goal and shortest path. How to take care of that, if the predefined trajectory is slighty curved. $\endgroup$ – Nag Feb 21 '18 at 15:07
  • $\begingroup$ Exactly this is the focus of elastic bands. It focuses on generalized path with local deformations. Distance is a factor, but distance to path not distance to current position $\endgroup$ – 50k4 Feb 21 '18 at 21:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.