create a pedestrian environment, where the pedestrian is rewarded for walking as far as possible from the spawn point.

[ll7]

Create a gym environment where the pedestrian is rewarded to walk longer distances.

Action space:

• walk direction
• walk speed

observation space

• segmentation camera
• init position
• current position
• current orientation

reward:

only based on the distance from the init position

expected issues:

• action space defintion for a walker
• input layer definition for a camera image and float numbers

[ll7]

carla-test/carla_env.py already has a first implementation of CarlaWalkerEnv, but I want to seperate this into a new file.

walker-env/carla_walker_env.py has the new environment definition

[ll7]

It is currently impossible to get the maximum size of the map and the start position would be a static observation during each episode. Therefore, I propose we use the distance from spawn point as an observation. Distance as 2D or 3D vector with a sign.

[ll7]