Carla Simulator

Tgus repository is original from a Zhihu blog:

--Carla Learning Tutorial

Direct Blog Links

Tutorial 1: An Overview of the Carla Concept
Tutorial 2: Install Carla
Tutorial 3: Use Basic API
Tutorial 4: Synchronize Mode
Tutorial 5: Traffic Manager
Tutorial 6: Behavior Planning(Part1)
Tutorial 7: Behavior Planning(Part2)

Installing Carla

Make sure your device has at least 100G free space, the unreal engine takes a lot of space.

Using Carla for TC-MAPF simulation

Path data

Using crossroad_waypoints.py to fit the discrete waypoints of TC-CBS output. The data are stored in python/path_data. Be careful because I aligned the map coordinate in the Carla city to the coordinate of our grid map.

In each case, for example in iros_final_paper_demo/test4AgentsSmall_straight2teams, put the original txt files of path results in it. And run crossroad_waypoints.py will save to the /fit folder.

Run Simulation

Using example_cross_xxxxx.py to run the simulation. In Carla simulator, press the play button and then run the code in the terminal. basic_api.py provides a very simple introduction of running the simualtion, where you can enable or disable any building or element in the city envrionment. You can set the weather, and even plot the planned path on the city. These examples should be self-contained. Following them will help you know how to make a environment and make cars move.

Make sure to refer to Carla's official document. The key is to understand how to spawn a vehicle, and then set its pose, which is easy.