Supernovae

Hi @bigsuperZZZX, as mentioned above, I am attaching a video demonstrating the problem mentioned above. as seen in the video, the drone moves into the obstacle, and checkCollisionCallback was not...

> Try to set a larger final_cost threshold. okay, I can do that, but I am willing to understand why this is happening, why checkcollisioncallback is not able to detect...

to isolate the problem, I tried to recreate the issue in the original implementation of main_ws. and I am seeing similar results as seen in the image below, the drone...

> I think that the planner failed to plan a new trajectory when the current one ran into an obstacle because the cost of the newly planned trajectory exceeded your...

Hi @bigsuperZZZX, any update on this?

Hi @bigsuperZZZX, earlier I thought the problem is caused because of the resolution, but I think that's not the case, since it's a discretized problem with consideration of the resolution,...

> It's really unclear to see the relative position between the trajectory and obstacles. Please modify the visualization parameters to make the obstacles more clear, like this:  Hi @bigsuperZZZX,...

> The planner plans trajectories at a given frequency like 1Hz, so a new trajectory is always generated before the drone reaches the last 1/3 part of the previous trajectory....

I have recorded a video, I tried to mimic the config you mentioned, I hope this video is more clear https://user-images.githubusercontent.com/51359628/217457806-4da6dd42-c013-4b97-bb05-ff7828b696d4.mp4

> I think that the cost threshold is too low. yes, you are right, but to test the worst-case scenario, I have put the cost too low.