Loading
Cookies help us deliver our services. By using our services, you agree to our use of cookies. Learn more

crowdAI is shutting down - please read our blog post for more information

NeurIPS 2018: AI for Prosthetics Challenge

Reinforcement learning with musculoskeletal models


Completed
4575
Submissions
477
Participants
65082
Views

Reward for Round 2

Posted by Yongjin over 1 year ago

The target velocity should be from ‘prev_state_desc’ instead of ‘state_desc’. I think that a target velocity should be given when an agent takes an action. Could you check the code from ‘reward_round2()’ whether it is intentional or a bug?

Big penalty for not matching the vector on the X,Z projection.

# No penalty for the vertical axis penalty += (state_desc[“body_vel”][“pelvis”][0] - state_desc[“target_vel”][0])2 penalty += (state_desc[“body_vel”][“pelvis”][2] - state_desc[“target_vel”][2])2

1