Team WPI-CMU: Darpa Robotics Challenge

DARPA Robotics Challenge Finals June 2015

The WPI-CMU robot (Warner) was the only robot that scored points (14/16 over two days), tried all tasks, and didn't fall or need physical human intervention (a reset). Our safety code was able to handle a bad step on the cinderblock rough terrain on day 1, and a wrist failure on day 2 which resulted in the robot almost pushing itself over. We had trouble with the drill on day 1 (human programming error) and due to the wrist hardware failure on day 2.

I am so happy and proud NOT to be included in this DARPA highlight video: DARPA's A Celebration of Risk (a.k.a., Robots Take a Spill)

I am also thrilled NOT to be included in these videos:

We are in this video, but not falling down: another somewhat funny video

WPI web page


Rubble, more, more

Push recovery during walking

Sagittal push recovery (marching in place)

Robust lateral balance (marching in place), more

Faster walking

Time lapse video of DARPA Robotics Challenge Day 2

Time lapse video of DARPA Robotics Challenge Day 1

Time lapse video of practice for DRC Finals

My stress relief video for the DRC Finals

DRC Trials 2013

DRC Trials Preview 2013 Shows ladder climbing and Ironman heart.


Humanoids 2015 Panel: Lessons Learned

Dynamic Walking July 24 2015

Talk at CMU: DRC Debrief July 9, 2015. Powerpoint only
Powerpoint and videos 0.5GB, Almost all the videos are on this page already.

Papers and Theses

Robust dynamic walking using online foot step optimization, S Feng, X Xinjilefu, CG Atkeson, J Kim Intelligent Robots and Systems (IROS), 2016 IEEE/RSJ International, 2016.

A distributed MEMS gyro network for joint velocity estimation, Robotics and Automation (ICRA), 2016 IEEE International Conference on, 1879-1884, 2016

WPI-CMU lessons learned paper: What Happened in the DARPA Robotics Challenge, and Why? submitted to the DRC Finals Special Issue of the Journal of Field Robotics.

WPI-CMU DRC Finals overview paper: NO FALLS, NO RESETS: Reliable Humanoid Behavior in the DARPA Robotics Challenge Humanoids 2015.

Paper on control: Optimization Based Controller Design and Implementation for the Atlas Robot in the DARPA Robotics Challenge Finals Humanoids 2015.
Video on control Humanoids 2015.

Paper on state estimation: Center of Mass Estimator for Humanoids and its Application in Modelling Error Compensation, Fall Detection and Prevention Humanoids 2015.
Video on state estimation Humanoids 2015.

Paper on getting out of the car: Full-body Motion Planning and Control for the Car Egress Task of the DARPA Robotics Challenge Humanoids 2015.
Video on getting out of the car Humanoids 2015.

Paper on doing the door task: Human-Supervised Control of the ATLAS Humanoid Robot for Traversing Doors Humanoids 2015.

Paper on doing the drill task: Improving Robustness in Complex Tasks for a Supervisor Operated Humanoid submitted to Humanoids 2015.

Thesis on state estimation: Xinjilefu: State Estimation for Humanoid Robots

Thesis on control: Siyuan Feng: Online Hierarchical Optimization for Humanoid Control
Video: part 1, part 2, part 3, and part 4. The most interesting part is the question period starting in part 3 at 6:00 and continues through part 4.

An analysis of what happened at the DRC finals.


Robô é chutado e se mantém de pé, sem perder o equilíbrio; veja (Globo (Brazil); in Portuguese).