search

TUMFTM / global_racetrajectory_optimization

This repository contains multiple approaches for generating global racetrajectories.
GNU Lesser General Public License v3.0
473 stars 194 forks source link

Error when the spline_approximation.py function is called in the scope of prep_track.py #11

Open pedrocalorio opened 2 years ago

pedrocalorio commented 2 years ago

I was trying to use this API to get the track yaw angle and curvature based on GPS data to run my laptime sims.

Then I was trying to use your algorithm to understand how you import the track data and get the racing line from it. However, when I try to run the main_globaltraj.py I have this error.

It seems that there is a optimization function being called in line 95 of the spline_approximation function, and the data size of the arguments are wrong.

This was the very same error I had when trying to use only the functions I needed to calculate the curvature analytically.

Here is the error log:

Traceback (most recent call last): File "/home/pedroc/dev/global_racetrajectory_optimization/main_globaltraj.py", line 243, in helper_funcs_glob.src.prep_track.prep_track(reftrack_imp=reftrack_imp, File "/home/pedroc/dev/global_racetrajectory_optimization/helper_funcs_glob/src/prep_track.py", line 39, in prep_track reftrack_interp = tph.spline_approximation. \ File "/home/pedroc/dev/global_racetrajectory_optimization/venv/lib/python3.10/site-packages/trajectory_planning_helpers/spline_approximation.py", line 95, in spline_approximation closest_t_glob_cl[i] = optimize.fmin(dist_to_p, File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/optimize/_optimize.py", line 622, in fmin res = _minimize_neldermead(func, x0, args, callback=callback, *opts) File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/optimize/_optimize.py", line 772, in _minimize_neldermead fsim[k] = func(sim[k]) File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/optimize/_optimize.py", line 496, in function_wrapper fx = function(np.copy(x), (wrapper_args + args)) File "/home/pedroc/dev/global_racetrajectory_optimization/venv/lib/python3.10/site-packages/trajectory_planning_helpers/spline_approximation.py", line 146, in dist_to_p return spatial.distance.euclidean(p, s) File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/spatial/distance.py", line 518, in euclidean return minkowski(u, v, p=2, w=w) File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/spatial/distance.py", line 462, in minkowski v = _validate_vector(v) File "/home/pedroc/.local/lib/python3.10/site-packages/scipy/spatial/distance.py", line 301, in _validate_vector raise ValueError("Input vector should be 1-D.") ValueError: Input vector should be 1-D.

Process finished with exit code 1

rs38 commented 1 year ago

did you find a reason/solution for that error?

    s = interpolate.splev(t_glob, path) # s will be an array of arrays
    return spatial.distance.euclidean(p, s) #1-D error for s

https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.euclidean.html but this is too old : https://docs.scipy.org/doc/scipy/release/0.11.0-notes.html#minor-change-in-behavior-of-splev

ahmad12hamdan99 commented 7 months ago

I had the sam error, I solved it by installing the same versions mentiond in the requirements file.