Wings, Wheels, And Walkers That Move Humanity Forward

Rise to the challenge of building Wings, Wheels, and Walkers. Today, we begin the search for things that move and make the world a little bit better place. This is the first day of a new round in the 2017 Hackaday Prize and your renewed opportunity to show us what you’ve got.

We just closed off the IuT ! IoT round, a more inward focused challenge which called for builds that added meaningful connectivity to devices in our lives. With the Wings, Wheels, and Walkers challenge we turn our gaze outward to see what you can do build that really moves.

There is so much that falls into this category; personal transport, robotic assist, automated delivery, airborne agriculture — anything that moves or supports movement. Many of the finalists and winners from the past few years fall into this category. In 2015 the Light Utility Electric Vehicle won 3rd place, and of course the grand prize winner that year was a wheelchair-based system. In 2016 we saw a shoreline debris clearing robot and a modular robot system took the top spot. Now we want to see even more creations that move humanity forward.

The Hackaday Prize is a global engineering initiative that seeks out ideas and creations that have the power to do social good. Show off your creation and you’ve already accomplished that and inspired others to do the same. Many of the entries will be recognized beyond that. This year’s cash prizes total more than $250,000. Just for this challenge (which ends on July 24th) we’ll award 20 entries $1000 each. At the end of all six rounds, 6 of the 120 finalists will be selected to receive $50k, $30k, $20k, $15k, $10k, and $5k. Enter now!

Check out all of the entries so far, and keep your on Hackaday to find out the twenty finalists from the IuT ! IoT round, an announcement due in about a week.

13 thoughts on “Wings, Wheels, And Walkers That Move Humanity Forward

  1. I find it amazing how much more difficult it has been to make walking robots than anyone predicted. 30 years ago everyone was sure we’d have mastered it by now, but it’s 2017 and we’re largely still stuck with jerky shuffling at 1 mile an hour. If only motors could surmount the torque-speed tradeoff that evolution so elegantly solved with muscles.

    1. There’s a huge difference between jerky shuffling with a statically stable gait on a flat floor, and jerky shuffling with a dynamically stable gait on rough, loose terrain with fast enough and accurate enough reactions to recover from being tripped up.

      Muscles sure do have an impressive blend of strength, speed, power, efficiency and accuracy though.

      1. It’s all in the feedback loop.

        There are several layers of feedback to detect your limb position, speed, tension etc. and predict the amount of force you need to apply based on cognitive processes – i.e. whether the milk carton is full or not. People make the same mistakes and jerky motions as the robots when the prediction goes wrong, which points out to the source of the problem: the machine is not intelligent enough to learn how to apply its strenght.

        For another example, take a well-rehearsed excavator driver. The motions of the machine become fluid and accurate as if it was their own arm, to the point where they can open a coke bottle with a ten ton machine for a trick.

    2. It’s not really the motors, but the software. Nobody wants to tackle the problem of passive dynamic walking, because it’s difficult, and it takes more computing to predict where you end up than what they’re willing to spend. In other words, it’s really a problem of (the lack of) AI and understanding of the dynamics of a walking system, rather than a lack of suitable hardware.

      Some have already built mechanical walking machines that step down a grade, or run with a simple clockwork mechanism and spend nearly zero energy doing that, but for some reason all the robotics researchers want to cling to the same old “keep your center of mass and all the force vectors pointing to within your base of support” thing that they -know- how to compute and implement, which then leads to robots that use 4x the energy to move around and they walk with their knees bent like they had just shat their pants.

      1. Click bait time! no, but some passive dynamic walkers and others:

        History:
        Collins passive dynamic walker
        https://www.youtube.com/watch?v=AdK7opXqoro
        McGeer passive dynamic walker
        https://www.youtube.com/watch?v=WOPED7I5Lac
        Cornell ranger
        https://www.youtube.com/watch?v=KLepY1AsaRk
        Atrias
        https://www.youtube.com/watch?v=dl7KUUVHC-M
        Cassie
        https://www.youtube.com/watch?v=CVcfNBYBYqA

        If you want to learn more go through this course:
        https://www.edx.org/course/underactuated-robotics-mitx-6-832x-0
        Textbook:
        http://underactuated.csail.mit.edu/underactuated.html
        Framework:
        http://drake.mit.edu/

        Also take a look at Gazebo and ROS, I am currently working on a workbench in FreeCAD to be able to export SDF/URDF for fast design/simulation iterations, my goal is to build a framework for a actuated biped walker based on passive dynamic walkers.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.