Edouard Leurent

Yes it is possible: the `action_type` is only relevant for the controlled vehicle(s) (or ego-vehicle). Other (uncontrolled) vehicles in the scenes do not follow this action type. In the environment,...

I don't know, here the vehicle is overlapping with the goal location, but this case should to be handled by the np.linalg.norm(v.position - self.goal.position) < 5 case Not sure why...

> Let's say I'm in the middle lane, how can I get the index of the left and right lanes and the current lane first You can access the ego-vehicle's...

Hi, This is because this config["observation"]["features_range"]["vx"/"vy"] does not control the range of speeds that can be reached by the vehicle, but only how the observed speed values get normalised into...

Yes I think this is supported already, you can configure the `highway-v0` env with some >1 `controlled_vehicles`: ``` import gymnasium env = gymnasium.make('higway-v0', config={'controlled_vehicles': 3}) ``` See https://github.com/Farama-Foundation/HighwayEnv/blob/81c11d6ff50a4fecf05a8341edaecfc245924e0b/highway_env/envs/highway_env.py#L35C18-L35C37

There is the [`parking-v0` environment](https://github.com/Farama-Foundation/HighwayEnv/blob/master/highway_env/envs/parking_env.py) which is a bit similar, where the task is to reach a goal position, heading, and speed(=0). But it is not a multi-agent environment right...

Hi! It's definitely possible, you should look into SB3's `VecEnv` see [documentation](https://stable-baselines3.readthedocs.io/en/master/guide/vec_envs.html), they have some examples. Essentially you have to use `make_vec_env`, and make sure you are using the `SubprocVecEnv`...

You should edit the environment definition, and in particular the part where the goal is defined: https://github.com/Farama-Foundation/HighwayEnv/blob/81c11d6ff50a4fecf05a8341edaecfc245924e0b/highway_env/envs/parking_env.py#L193 By default, parking spots are defined as small lane segments and we use...

> If I only make one side of adjacent parking spaces coincide with each other through calculation, there will always be a line that does not meet the requirements at...

not sure why you're adding -pi/3, if you want the vehicle to be aligned with the parking spot you should use the lane heading as target heading, so `Landmark(..., heading=goal_lane.heading)`