search

SigmaEmbeddedEngineering-eg / Real2Sim_gazebo

One of the main problems in AD/ADAS simulations in Automotive industry is that simulation is too much ideal and far from realistic and artifacts representation in different ADAS systems. this repo demonstrating a method of using DNN, to represent LiDAR sensor model in simulation environments
Apache License 2.0
15 stars 4 forks source link

How to use this repository to simulate LiDARs in Carla #1

Closed amansrf closed 1 year ago

amansrf commented 1 year ago

Are there any specific instructions on how this repo could be used with Carla? I am trying to simulate Luminar's Hydra LiDAR in Carla.

khaledelmadawi commented 1 year ago

1st you need to train the network on real luminar traces and replace the path of the DNN to the new path: https://github.com/SigmaEmbeddedEngineering-eg/Real2Sim_gazebo/blob/master/real2sim_lidar_sensor/models/hatchback_ego/model.sdf#L124

2nd you need to reconfigure the horizontal and vertical configurations to the resolution of ur lidar sensor: https://github.com/SigmaEmbeddedEngineering-eg/Real2Sim_gazebo/blob/master/real2sim_lidar_sensor/models/hatchback_ego/model.sdf#L92 https://github.com/SigmaEmbeddedEngineering-eg/Real2Sim_gazebo/blob/master/real2sim_lidar_sensor/models/hatchback_ego/model.sdf#L98 3rd as per my knowledge Luminar has different artifacts which require some preprocessing, you will need to add them. we can have a short call to share ideas, you can contact me @ khalid.elmadawi@sigma.se