Feature: Fastlio working with Go2 navigation stack

[kahzmic]

Problem

I've integrated the Mid-360 LiDAR into the current Go2 navigation stack, using the Fastlio native module. I also added a Fastlio test case that replays point cloud data to the voxel mapper with Rerun visualization.

Solution

This integration includes a module that transforms the point cloud from the Fastlio native module to the Go2's base frame accounting for the sensor's physical mounting position.

Breaking Changes

None

How to Test

• Follow the docs/usage/sensors/mid360.md to setup the Mid-360
• run dimos run unitree-go2-fastlio
• If you don't have access to hardware you can run a replay of the blueprint that will show you the costmap generated by the Mid-360 on the Go2
• To run the test pytest dimos/hardware/sensors/lidar/fastlio2/test_fastlio_column_carving.py -v -s -m integration

Note: When running this test case you may run into an error where a thread notices that your network buffers are small and attempts to prompt you so that it can run some sudo commands. If this happens run the block of commands below.

# Enable multicast routing for local testing
sudo ip link set lo multicast on
sudo ip route add 224.0.0.0/4 dev lo
# Increase UDP network buffers
sudo sysctl -w net.core.rmem_max=67108864
sudo sysctl -w net.core.rmem_default=67108864

pytest dimos/hardware/sensors/lidar/fastlio2/test_fastlio_column_carving.py -v -s -m integration

Contributor License Agreement

• [ x] I have read and approvehardware

Contributor License Agreement

• [ x] I have read and approved the CLA.

[leshy]

those modules autobuild now, just need to deploy, no need for manual building