I have taken the plunge and I’ve gone part time at work, dropping to three days a week, to give me far more time to concentrate on my lifelong passion of robotics with the aim of making it my career! This wasn’t exactly planned if I’m honest but the result of a recent epiphany, building on my experience with MacFeegle Prime at PiWars.
My plan is to use this time to continue learning ROS, finish development on MacFeegle Prime and it’s controllers, and restart development of a few stale projects. I’ll be making all of these projects open source too, to give back to the community that has, and continues to, inspire me to bigger things.
The projects should be familiar to long time followers of my blog and other channels, though as I’ll be rebuilding them to share they’ll have more sensible names. Sensible, but all puns…
This will be a resurrection of the first robot project I tried since my A-Levels, rebuilding my Original Roomba nicknamed “Bumblebee”. I was given this robot in a broken state by a friend of mine, it turned out that the issue was a duff battery so easy to get working again, it quickly developed a fault where it’s speaker died though and it lost its voice which is where the nickname comes from. I used it as designed for more years after that before it just stopped working, none of the buttons worked and it was totally unresponsive, looked like the main circuit board had died.
I made a series of videos showing the in’s and out’s of reverse engineering the robot culminating in being able to drive it around using a gamepad, I never recreated its autonomous functions of being able to hoover by itself though.
Project NE-1 will resurrect this project as a way to show how a cheap old hoover, which can be found on eBay for as little as £25, can be reborn using a Raspberry Pi and given new purpose. The newer versions have a diagnostics port which makes this even easier, but this will concentrate on how to reverse engineer the hardware and hopefully show how this or an RC car can be converted in to a very affordable robotics platform. I’m hoping that this will show that there are ways for anyone *cough* can get in to robotics.
NE-5 (MacFeegle Prime)
MacFeegle Prime was my entry for PiWars 2020, it won the public vote for fan favourite and came second overall! This robot has had a lot of blood, sweat, and tears put in to it but it suffers from a lot of legacy problems from early in its development, essentially I didn’t *really* know what I was doing, making it up as I went along and in some ways it shows.
The plan for MacFeegle Prime is to finish the build as per the MVP and get him to the basic level to compete in PiWars for when it’s held next year. At that point I’ll have learned how to build an awesome robot, more importantly though all the ways not to build a robot. Using all my experience, and the parts, from MacFeegle Prime I’ll design and build NE-5 which will be redesigned from the ground up. The mechanical files, BOM, and code for NE-5 will be published so anyone can build one to use for research or just for a laugh.
MacFeegle Prime has been build using off the shelf parts, some slightly modified, along with 3d printed custom components. It’s this combination of easy access to components and relatively simple construct that makes for an incredibly capable mobile robotics platform. Also a testbed for larger versions. In future phases there will be an NE-5L, and hopefully an XL, closer to the size of it’s inspiration…
NE-Where (The Luggage)
Has to be done!
The Luggage has been a daft project that has been on and off the back burner since EMF 2016, when I realised how annoying it was to walk back and forwards from my car to bring all my crap to the campsite. Fast forward a couple of years to EMF 2018 and Hacky Racers was born! I decided to resurrect the project as a racer, with a pair of 2KW motors and a lot of moxie! Again, I was making it up as I went and I didn’t have the experience with CAD I do now, I didn’t really have a plan so much as a pLn, left everything to the last minute and it didn’t come together at all.
After the first couple of events for Hacky Racers I realised I really like commentating and running the races, handy as everyone else wanted to race! This also meant that I didn’t have to worry about falling within the rules (which was already in a lot of grey areas) nor have to worry about getting it done “in time” for anything, which meant it didn’t get done…
Fast forward to now and I have the majority of all the parts to not only finish it but make it much more capable. Rather than being a racer, then a racer that could be remote controlled, it’s now going to be a robot that I can ride.
To complement the relative simplicity of the NE-1 and the complexity of NE-5, NE-Where will be my heavy duty rover. I’ll be redesigning it from scratch along similar lines to my bike trailer. It was literally built around an 84L Really Useful Box, and that’s my plan for NE-Where too. As well as having a daft alter ego in the shape of The Luggage, legs and all, it’ll also have a full set of wheels for more sensible uses.
This will be used for research into logistics robots, rough terrain navigation, and as a base for NE-5XL too.
Robots aren’t much use if they can’t be controlled and for PiWars I made a custom controller for MacFeegle Prime based around a pair of three-axis joysticks, a bunch of switches, and a Raspberry Pi.
This controller uses ROS, as does the rest of the robot, to communicate and control it remotely. I got it to the point where it could drive the robot around and control it’s head too, with the live video stream I could control it remotely. The plan was to have a driving mode and manipulator mode, alas I didn’t get the arms integrated in time to get that far.
This was always intended to be reused for other projects, hence NE-Thing, and the design will be extended with that in mind. Along with WiFi and Bluetooth that are available on-board the Pi I’ll be adding an NRF24L01 transceiver as well as a GPS reciever. This will be useful for outdoor projects where WiFi isn’t available and GPS will allow for dynamic return to base control too.
In this phase this controller will be used for the robots above, longer term I intend to build a drone too where the return to base functionality should come in to its own.
One thing I’m going to research and develop further is teleoperation, with NE-5 having stereo vision I wanted to build a waldo controller to allow for more intuitive control over the robot’s arms.
The design above was a pretty quick design I threw together in a few hours, that I can say that alone shows how far I’ve come, and needs a lot of improvements to be actually useable. Likely I’ll switch to encoders of some kind instead of potentiometers but that needs researching. Due to the sheer number of potentiometers involved this will be based around either a Teensy or Arduino Mega, I think it’s 16 joints in total after all.
At the minute I’m not sure if this will connect directly with the robot or be a peripheral of the NE-Thing, I’ll have to make that call when I get around to it.
Dropping to three days a week means I’ve still got a regular income and I’ll be able to cover my bills, I’m incredibly lucky to be able to say that, and the vast majority of parts for all the projects above I’ve purchased over the years already. For the foreseeable future I’ve plenty to be getting on with and all the parts I need, but for future phases I’ll need to find that budget from somewhere. My plan is to launch a Patreon page to help fund future work with the hope of taking this full time too.
If you stuck around to the end, thank you! Let’s see where this ride takes us!