Trajectory Force Field Implementation

LCAS / iliad

L-CAS specific works on ILIAD project

6 stars 5 forks source link

Trajectory Force Field Implementation #25

Open marc-hanheide opened 5 years ago

marc-hanheide commented 5 years ago

In order to allow local trajectory changes we look at a force field model, that takes local obstacles and uses the reference trajectories, with some local curvature constraints observed (i.e. don't deviate too much from reference)

look for relevant papers and report on them
very hacky prototypical implementation in jupyter

marc-hanheide commented 5 years ago

see if we can force teb to do this for us, or even the steering functions?

MFernandezCarmona commented 5 years ago

We can rule out steering functions, as they generate trajectories that are compatible with ackermann like vehicles. Looking at TEB

MFernandezCarmona commented 5 years ago

Ok. TEB uses some sort of graph optimization engine. But I found more information about trajectory optimisation. I have implemented a really simple optimization script in python. @marc-hanheide , have a look at it in the jupyter notebook here: https://lcas.lincoln.ac.uk/user/mfernandezcarmona/lab/tree/work/Trajectory_optimization.ipynb

MFernandezCarmona commented 4 years ago

This could be introduced in https://github.com/LCAS/iliad/blob/master/dynamic_constraints/scripts/taskManipulation_node.py in order to alter local trajectories.