When given the opportunity to build a robot for his systems engineering assignment, Warren thought big and made bigger.
In the past, Makers would need to reach for the Meccano or get out the hand saw and woodworking tools to make their large scale robot designs. Thanks to laser cutters becoming a lot more affordable, Makers are now able to create amazing large-scale robots in their own home or classroom.
That was the case for year 12 student, Warren Rogan who has built a human-sized robot out of laser-cut timber as part of a systems engineering subject.
Read how Warren build his robot, aptly named ATOM, and used Arduino technology and sensors to make the robot mimic his arm movements.
Thank you for submitting your project to us to feature, Warren. First, tell us a little about yourself.
I’m an eighteen year old that lives down on the Mornington Peninsula (South of Melbourne) and I am currently in year 12 attending Mornington Secondary College. I enjoy gaming and building things, hence the ATOM robot (i.e. from the movie Real Steel).
Your ATOM robot looks amazing. What got you into robotics?
I got into robotics after getting interested in electronics as part of design technology at school where I used a laser cutter and 3D printer. Basically, from there I chose electronics as an elective subject, which became a pathway to systems engineering in VCE where I learnt about Arduino and servos. This was pretty much my foundation of basic robotics knowledge.
We’re pleased to hear electronics being taught at your high school. Was ATOM part of your Systems Engineering class and was it just on your own or a team project?
I built ATOM last year when I was in year 11, which was a year 12 systems engineering class because I’d accelerated the subject a year early. Systems engineering in year 12 is basically all about creating a project that has electrical and mechanical systems, and each student chooses an individual project they work on.
You obviously have a talent in robotics if you’re able to do year 12 curriculum in year 11. Well done! Was this task assigned to you or was there other inspiration for why you chose to design this project?
I chose this project as I am a Marvel fan and thought that making something like the Iron Man armour would be great. However, I soon made it more of a robot rather than exosuit after realising that it would be much easier to do that way. I designed it to mimic human movements rather than being autonomous as I had already learnt how to program a servo to mimic a potentiometer as part of my year 11 systems. I wanted it to be similar to the iron man suit in Iron man 3 and the “shadow function” of the ATOM robot from Real Steel.
Awesome! It’s great that Makers can build robots based on inspiration from movies like these. What was the reaction from your robotics teacher?
When I explained my idea my teacher was very supportive and helpful, even though he had some misgivings about the size of the robot itself.
That’s quite evident in the photo of you standing beside your robot. ATOM is huge! Can you tell us how it works?
Basically, my robot mimics my movements by reading the analogue values of a potentiometer attached to one of my arm joints. It maps those values against the minimum and maximum value of the analogue signal from the potentiometer that I recorded in some tests on the analogue read function. It then converts it to a value between 0 and 180 degrees and writes this value to the appropriate servo in the robot.
That sounds straightforward, but we’re sure it took a lot of time to get it right. Does it hold its own weight and is it able to walk or is it just arms that move?
Unfortunately, although there are servos in the legs of the robot, it cannot support its own weight as the servos are not powerful enough to counter the torque due to the mass of the upper body and arms. For this reason, it cannot hold its own weight or walk and can only use its arms, neck and eye.
There’s a challenge for you in the future perhaps. What kind of sensors and how many sensors are you using?
There are five functional potentiometer analogue sensors and three push button switches that I use to control the robot. Four of the potentiometers are attached to my wrist and elbow joints and the final functional potentiometer on the back of my hand to control the neck servo. The three buttons are used to turn the RGB LED in the robot’s head on and off.
I added the RGB LED into the head, largely to give the robot some character, as otherwise it seemed a bit plain. I chose RGB rather than a single colour LED pretty much just for the coolness factor. Silly I know.
Adding the RGB LED is a nice touch. Were there a particular servo type you needed to use and does it use commonly available parts?
It uses regular parts, aside from the laser cut pieces. It uses a generic Mega 2560 board as the controller, and regular blue servos (9G) and higher power 13kgcm servos for the major joints. The potentiometers are all regular 100K ohm and I also used two 170 pin breadboards to make it easier to wire the servos and sensors.
Sounds like many of our readers could build a robot for themselves. Why did you choose to use a Mega 2560 controller and what functions does it perform?
I wanted a single board that could control the most servos without the need for a shield. The Mega functions as the brain of the project and converts all the sensor signals to the PWM for the servos and toggles the I/O pins connected to the RGB LED. The potentiometers register how much of an angle I have moved my arm, relative to the attached joint and send an analogue signal to the Mega chip.
Tell us more about the laser cut skeleton. How did you go about the design and getting the plywood cut?
I designed the laser cut parts using Illustrator by creating boxes that would mount together to create the form of a person. I then laser cut the bigger pieces at school and then used my generic K40 laser at home for the smaller parts from poplar plywood. The plywood is light but relatively strong.
You have a laser cutter at home? That’s great! What were the main challenges you needed to overcome in the design and build?
One of the main challenges was the total cost of the project, as the servos and Mega board were quite expensive. The servos alone cost almost $200, and after seeing that the parts were cheaper internationally but would take a month to arrive. That’s when I decided that I would start my own online store called Milliamp Diode (https://milliampdiode.com), so that both schools and hobbyists could get a better price for their electronics.
Sounds like you are a young entrepreneur in the making. Were there any other challenges?
Another challenge I faced was how to attach the potentiometers to my elbow and wrists. Keeping them in place would be difficult if I attached them to clothing, but they needed multiple points of contact to function accurately. I eventually worked out a way to create a multi part harness that kept the potentiometers in place but could be adjusted and easily removed and put on.
We need to get creative sometimes to find the best solution. If you were to start again, what would you do differently?
I would organise the wiring connections better, as honestly the number of times that I was testing the system and it wouldn’t work due to a wire popping out was ridiculous. For that reason, I would use more pin headers and custom sockets.
Wires consistently falling out would have been annoying and time consuming to fix. Apart from designing a wiring harness, do you have plans to extend your robot’s capabilities?
I intend to continue to develop and redesign the robot. I am currently working on Mark 2 of the robot, which currently has more wire organisation and more reinforcement on the waist joint. I also intend to use stronger servos and design the pieces to cause less stress on the servo heads, as I believe some of the servos were damaged because of this.
That’s a costly exercise to replace damaged servos. Perhaps you need to research methods that other robot designers use to help avoid the entire load being on the servo itself. What would your advice be for any of our readers who want to make a similar robot for themselves?
I would recommend that they design the robot to be strong but light, so that the servos are not overloaded and perhaps integrate a counterweight, gyroscope or accelerometer to help keep the robot steady.
Have a plan on what you want to do and don’t be afraid to redesign parts if they don’t work as they should.Sometimes that simple redesign can make the entire project fit together and function much better.
A self-balancing robot would be cool. Thank you for sharing your robot project with us. We look forward to seeing your future builds.
Image credit: VCAA