Info

La pregunta está cerrada. Vuélvala a abrir para editarla o responderla.

Need Help with Curvature spikes

16 visualizaciones (últimos 30 días)
Heavenly
Heavenly el 6 de Dic. de 2025 a las 13:22
Cerrada: John D'Errico el 6 de Dic. de 2025 a las 16:04
Hello everyone,
I have been working on UAV guidance and right now working on writing the code for obstacle avoidance. When I plot the curvature of the path that I have planned, the curvature spikes at two places during the start of the avoiding of the obstacle and the end too. Can somone please help me with this?
  2 comentarios
John D'Errico
John D'Errico el 6 de Dic. de 2025 a las 16:01
Editada: John D'Errico el 6 de Dic. de 2025 a las 16:04
There is no question about MATLAB in this, as opposed to a general plea for help on a vague question.
If you have a problem with the curvature, then break it down into simple sub-probems that you can handle.
  1. Learn how to estimate the curvature of a path, and to find the maximum curvature.
  2. Parameterize the probem, so that you can choose different paths.
  3. Search over the set of various paths, minimizing the maximum curvature. Choose the path that makes you most happy in whatever sense that can be here. Only you know how to define goodness.
But when you actually have a question about MATLAB, then ask it.
Sam Chak
Sam Chak el 6 de Dic. de 2025 a las 16:03
I can't help but only guess about what may have occurred. Based on your illustrative description, the spikes are likely caused by significant repulsive forces generated while attempting to conservatively avoid obstacles, thereby ensuring a safety buffer against worst-case scenarios.
If the repulsive force is derived from clearly defined mathematical equations designed to avoid obstacles, you may consider posting the MATLAB code so that path planning experts in this forum can examine it.
However, if the repulsive force is produced by a seemingly learning-guided black box system, where you can only see what goes in (input) and what comes out (output), without knowing about the internal workings of collision avoidance rules, then you will need to make the black box adapt and learn from failures in simulation environments.

Respuestas (0)

La pregunta está cerrada.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by