I don’t think it’s a stretch to say that we are in a golden age for robotics. A lot of technology available in the realm of hobby robotics didn’t exist 10 years ago, or didn’t exist at a price that most hobbyists could afford.
Related, and just as important, is how much information is currently available. I’m a mechanical engineer by training, and so most electronics and programming is a challenge for me (part of why I’m doing this project is for me to learn those skills).
A lot of this information is available on the internet, and if it’s not, there’s usually a genius lurking on some website who has solved a very similar problem and is willing to lend a hand (if you ask nicely and do a minimum of homework before asking).
So before I get into the specifics of the Mower Project, I want to list several of the “Giants” whose shoulders I’m standing on. These are folks who have convinced me the Mower Project is possible, inspired me to do something that I know will be difficult, or have solved some epic technical challenges I foresee with the project.
Rusty Nails Workshop
Alligator over at the Rusty Nails Workshop developed something very similar to what I had in mind for the Mower Project back in 2014. The Mower Project was a pipe dream until I saw his blog post about his autonomous lawn mower.
His system uses Swift Navigation’s Piksi version 1, a Sabertooth 2X32 motor controller, and a laptop running a python program for automation as I understand it. He had a camera mounted on the robot for vision.
I think Alligator intended to implement vision based automation on it with that kind of setup. I’m not sure what happened to his project, but it looks like he wasn’t getting good results with the Piksi version 1 and decided to move the hardware over to an epic RC snow plow robot.
I personally think he jumped into making the mower portion of his design too soon. I would have spent more time perfecting the automation aspect before implementing the mower portion, but you can’t argue with his get ‘er done attitude. He accomplished more in one year than I have in four.
Paul Breed is a mad scientist. I mean that in a good way. I’ve read his blog Unreasonable Rocket from start to finish, and what is most inspiring to me is how it seems like one day he just decided he wanted to build rockets. And so he did.
Paul basically wrote his own Arducopter code before Arducopter existed. He then went on to use that code to design an auto-stabilized rocket and placed 2nd in the 2009 X Prize competition.
Paul’s work doesn’t have much to do with the Mower Project, but seeing a guy basically build a hovering rocket from scratch blows my mind. When I get discouraged about the Mower Project, I go read some of Paul’s blog. He’s had rockets blow up on him during test runs. Compared to that, my little wiring issues are chump change.
Speaking of Ardupilot, I would be remiss if I didn’t mention this awesome open source autopilot software. These guys have basically done the hard work of programming a quadcopter to fly autonomously.
I am very grateful for the folks that fly quadcopters and develop the Arducopter portion of their software, because a lot of awesome features get ported over to the Ardurover platform.
The combination of the Pixhawk flight controller and Ardurover is what I intend to use for the Mower Project. This hardware and software combination is used by guys that build things like autonomous tractors, vertical takeoff aircraft, and aerial mapping quadcopters.
Deep South Robotics
Robby over at Deep South Robotics is another genius I’ve come across in my research. He wrote an article about implementing an RTK GPS receiver a while back that received a lot of interest from folks like myself attempting to automate an AGV in a high multi-path environment.
Robby has a lot of good information about AGV construction best practices. He has done a great job highlighting some cool robotics solutions (like these Gore vents) that I was previously unaware of.
I’m not sure why he stopped posting to his blog, but I suspect some top secret government program made him an offer he couldn’t refuse to work on some next generation defense project.
There’s a picture of him on a large riding mower with his robotic platform installed over at his blog. I’m not sure he ever mentioned it on the blog, but I secretly think he is a lot further along than us mere mortals at making a truly autonomous mower.
I’d also be remiss if I didn’t mention my wife who allows our office to be cluttered with random robot parts and allows me to use her for my rubber duck debugging conversations. Her support makes this project both fun and rewarding. I’m a tight-wad, and she has helped encourage me to spend a little money on the parts I needed to get the Mower Project rolling.
I expect this blog will serve as a place to document the Mower Project’s progress, and to motivate me to work on it. I’m hopeful it will also serve as a forum for folks working on similar projects to swap notes. Next time I’ll give an overview of the Mower Project and some of the hardware I’m using.