Hey there!
I am working on a project using a Jetson Orin Nano and a Cubepilot for an autonomous system in a drone. So far, Nvidia Isaac ROS seems to be the best option for all perception and exploration tasks. However, I noticed that Isaac ROS Nvblox has only one costmap layer. I would like to know if there is a way to obtain cost maps from different levels and use them in Nav2, so the drone can fly over them.
Hi @jim.fabian
Yes, Isaac. ROS Nav2 only generates one map for navigation.
To better understand your request, can you please explain why you need more than one costmap?
meanwhile I forward your request to the engineering.
Best,
Raffaello
1 Like
Hello @Raffaello, thank you for the response.
I am using a drone to navigate at different heights within a warehouse, specifically across various levels of the racks. My approach involves using different cost maps, one for each level. For instance, I navigate on one layer for a specific level while maintaining the height, then switch to a new cost map for the next rack level and navigate accordingly.
I am using this approach because 3D path planning is not yet available in ROS2. Reinforcement Learning is an option I am considering for later, once PX4 simulation is supported in Isaac SIM or Pegasus Sim supports uXRCE-DDS.
Hello @Raffaello , I was wondering if you have any update. I am still looking for options and Nav2 is the best one, but I still need multiple cost maps.
thanks!
No, unfortunately, the engineers are pretty busy these days.
I ask again for updates.
I keep you posted
Raffaello
1 Like