[find_and_avoid] Fix convergeness issues

aidudezzz / deepworlds

Examples and use cases using the deepbots framework (https://github.com/aidudezzz/deepbots) with the Webots robot simulator.

GNU General Public License v3.0

50 stars 23 forks source link

[find_and_avoid] Fix convergeness issues #41

Open eakirtas opened 3 years ago

eakirtas commented 3 years ago

Currently, agent of find and avoid seems that it is not converge

KelvinYang0320 commented 3 years ago

@ManosMagnus I have fixed the issue that robot stop moving after the first episode by adding two lines of code. However, it seems weird that there are several additional resets. Did you design this for some purposes? 20210716_174505

KelvinYang0320 commented 3 years ago

Currently, it seems to be able to converge. Link to My fork However, the robot doesn't reach the target when using agent.choose_action_test() instead of agent.choose_action_train()after the training is done. Still trying to figure out how to fix it.:thinking: