ChallengeRocket
  • Challenges
  • Find jobs
  • Candidate Hub
  • For business
  • Log in
  • Sign up now
Menu
  • Home
  • Challenges
  • NVIDIA® Jetson™ Developer Challenge

NVIDIA® Jetson™ Developer Challenge

NVIDIA® Jetson™ Developer Challenge
  • Winners announced
  • Winners announced
prize pool $42,789 gross

SEE RESULTS

SEE RESULTS

Oct 23, 2017 - Feb 18, 2018 23:59 GMT
ONLINE CREATIVE CHALLENGE
  • Challenge outline
  • Resources
  • Participants
  • Projects
  • FAQ
  • Results
  • Updates
  • Rules
NVIDIA® Jetson™ Developer Challenge
  • Challenge outline
  • Resources
  • Participants
  • Projects
  • FAQ
  • Results
  • Updates
  • Rules

YC

Yu-Ming Chen

Added: Feb 18, 2018

TAGS

  1. robotics,
  2. deep learning,
  3. reinforcement learning

TYPE OF PROJECT

Robotic Appication

WWW

hellochick.github.io/projects/Sim-to-Real/

VOTES: 3 LIKES: 2

Sim-to-Real Autonomous Robotic Control

  • play
  • Sim-to-Real Autonomous Robotic Control
  • Sim-to-Real Autonomous Robotic Control
  • Sim-to-Real Autonomous Robotic Control
  • Sim-to-Real Autonomous Robotic Control
  • Sim-to-Real Autonomous Robotic Control
  • pdf
  • pdf

    Project description

    Collecting training data from the physical world is usually time-consuming and even dangerous for fragile robots, and thus, recent advances in robot learning advocate the use of simulators as the training platform. Unfortunately, the reality gap between synthetic and real visual data prohibits direct migration of the models trained in virtual worlds to the real world. This project proposes a modular architecture for tackling the virtual-to-real problem. The proposed architecture separates the learning model into a perception module and a control policy module, and uses semantic image segmentation as the meta representation for relating these two modules. The perception module translates the perceived RGB image to semantic image segmentation. The control policy module is implemented as a deep reinforcement learning agent, which performs actions based on the translated image segmentation. Our architecture is evaluated in an obstacle avoidance task and a target following task. Experimental results show that our architecture significantly outperforms all of the baseline methods in both virtual and real environments, and demonstrates a faster learning curve than them. We also present a detailed analysis for a variety of variant configurations, and validate the transferability of our modular architecture. The architecture is implemented on an NVIDIA Jetson TX2 development board, and comprehensively evaluated on real robots.


    • previous project
    • next project
    By using our site, you agree to challengerocket's use of cookies in accordance with our privacy policy.
    ChallengeRocket
    Tech talent
    Challenges Blog Find jobs Candidate Hub
    Companies
    Business HR Blog Pricing
    Challengerocket
    FAQ EU Join Us Contact Us
    We connect Top Tech Talent with Companies. We are community of software developers, engineers, data scientists and tech lovers.
    Copyright © 2021 ChallengeRocket. All rights reserved.
    Privacy Terms and Conditions Service status

    Sign in

    Don’t have account?
    Create a new account
    or

    Forgot your password?

    Log in

    Forgot your password?

    Don’t have an account?  Start a free trial

    Create a new account

    Already have an account?
    Log in
    or
    • At least 8 characters
    • Uppercase Latin characters
    • Lowercase Latin characters
    • At least one number or symbol

    Reset your password

    Have an account? Log in Log in for business

    Start a free trial

    • At least 8 characters
    • Uppercase Latin characters
    • Lowercase Latin characters
    • At least one number or symbol