scout is coming [work-in-progress]
Updated: Jan 6, 2020
once again, why?
Yes, it's another robot dog! I did say we'd come back to the chip project, hence the name chipONE. Scout is the latest LARGE project I'm undertaking and this time I'm taking some lessons from the latest failure, chipONE. The problem with chipONE was a robot of that size needs to be DESIGNED not thrown together. I don't regret the time I spent on chipONE, it taught me a bunch of valuable lessons to take forward. But after spending a summer thinking about what was wrong with chip, I thought I'd have another go at the concept of a four-legged robot dog, this time with some design involved.
SCOUT is the next iteration of my robot dog project, and the goals for this one, I admit, are lofty. The plan, at the moment, is to make a four-legged robot with the following minimum capabilities to start with:
(1) Legged motion, simple walking and turning for now.
(2) Joystick, human control.
(3) Ardupilot-based autonomous outdoor navigation on relatively light terrain.
(4) ROS/Ardupilot/SLAM autonomous navigation of indoor spaces.
This project is for fun because I've always wanted a robot dog, but it's also for another reason. Legged robots are more-and-more looking like the reality of the future because of their ability to dynamically respond to the environment (being pushed they can stabilize, they can go up stairs, cross rough terrain, etc). But to get into the legged robot space, you can either blow 15-30k on spot mini, or similar robot, or you can go build one yourself. There are legged robot instructions out there, like Stanford's doggo, but most of the robots large enough to handle high payloads and real work are either custom platforms or way too expensive for an individual, startup, small business, or similar, to afford.
The SCOUT project is about making a robot dog large enough and capable enough to do real work using simple and robust engineering. The dog should have a decent payload capacity, it should be able to handle relatively rough terrain, it should be able to navigate on its own and relay data back to its operator. And it should be able to do so at a cost that doesn't break the bank because it isn't interested in being a product that makes profits.
We're starting the SCOUT project in the hopes that a few months from now we'll come up with a solution that's completely open-source that we can put on the web and will cost less than 4-6k to build. It will be highly capable and can be used for a number of industries including search and rescue, development, GIS and planning, nonprofit, and so much more. We want to make something that people can use as a development platform to get in to the legged robot space without having to destroy the bank, or having to rebuild something from scratch completely on their own. That's the goal of SCOUT.
I'll be posting regular updates on this blog, but the robot should be set to come out by the end of next summer 2020, if not earlier!
Excited for another cool, new project!