search

tud-airlab / mppi-isaac

Model Predictive Path Integral Control using isaacgym for rollouts, gpu-accelerated
178 stars 21 forks source link

Benchmark with local planner bench #20

Closed cpezzato closed 1 year ago

cpezzato commented 1 year ago

Use https://github.com/tud-amr/localPlannerBench for benchmarking with point robot, boxer, and panda.

maxspahn commented 1 year ago

We got first results for the point robot :rocket: results_comparison

It is with random obstacles and without the prior. I guess there is some tuning to be done, @cpezzato @eliatrevisan MPPI can do better than that.

cpezzato commented 1 year ago

Definitely the robots should be tuned better with mppi, the current versions are quite naive and bad! But awesome you've got the benchmarks 🎉🎉🎉

cpezzato commented 1 year ago

45 addresses this point