Building Thought-Wired BOT ver 0.0: Mechatronics Engineering Summer Internship Experience
Below is a blog post by our summer intern, Jono Wong. Jono is a final year mechatronics engineering student at the University of Auckland. He’s hoping to continue his studies and specialise in biomedical engineering.
While having the privilege to work at Thought-Wired this summer, I was asked to build a robotic vehicle that responds to inputs from the brain sensing headsets. Having already done most of the software aspects in 2015 while working with an mBot, moving onto the physical designs was very exciting. As soon as I was asked to do this task, my mind immediately raced from design to design. Moving forwards from imagining a Ferrari to imagining space ships.
However, this particular design was made to be a gift to our friend, Mandy, who has helped and continues to help Thought-Wired test and develop nous. As such, I was advised to create a design based on a truck her late father built. The truck resembled a 1931 Ford Pickup Truck with a red body coat and white fenders, and shiny silver tyre rims.
Before considering the designs, I first began by looking at the materials and facilities I was provided with:
1. Arduino Uno with a motor driver shield
2. Bluetooth module (HC-05)
3. Lithium-Ion battery pack
4. Motors and wheels
5. Laser cutting and 3D printing facilities
While doing this, I started to look for possible solutions and designs that can be made to accommodate what I was given. The main design concerns which I noted while looking at the materials were the Bluetooth unit which has yet to be configured, the unusually large battery pack that I was given, and the needless double axial on the motors as I am only running one wheel per motor. After finishing configuring the Bluetooth module and other preparations (like yanking one of the axials out of the motor), the design phase came next. Here, I had to design a fairly large and rigid chassis using 3mm MDF to accommodate the large battery whilst keeping in mind how the wiring to the Arduino unit will work.
The application of the laser cutter was very convenient as it was impressively fast and efficient; completing each print job from 5 to 10 minutes. In the event of mistakes, it was very easy to reprint and remake the pieces I needed. However, aside from the ease of printing pieces, the assembly of the chassis proved to be a frustrating task, as use of super glue often made a mess of everything (regrettably, also including my fingers).
Now I admit that it looks nothing like the truck, and that it perhaps looks more like a tractor than it does a truck. However, I was still determined to have it at least somewhat similar to the truck. As such, I decided to paint the robot to look like the truck, with red body coat and white rims. To do this, I bought some spray paints and primers from my local model shop, and began working on the robot. While I was painting, I noticed several issues which involved spray painting. To begin with, the coating on each layer had to be consistent. Excessive paint will cause bubbles and drip marks when drying, and this happened to primers as well. In this process, I would often accidentally apply too much paint in certain areas, causing me to perform touch ups with my paint brush. After several hours of waiting and several layers of paint, the final finished product looked something like this:
I wanted some special features on the robot aside from the headset functions, so I added in some holes at the front to allow LEDs to be installed. I decided to 3D print some conical covers for the LEDs at the front of the car. Aside from the conical covers, I also had to design and 3D print a ball bearing to allow a free spinning ball wheel at the front of the cart. This process took a long time, as I had to go through many iterations of designs to finally obtain a suitable ball bearing which had minimal friction with the ball for it to spin freely. While doing this, I came to the realisation that 3D printers, as cool as they sound, are actually really annoying. Without carefully considering the design, each iteration of printing takes anywhere between 40 minutes to an hour just to print a small sized object (about 5x5x5cm space). Besides that, the 3D printer also has tolerances which should be carefully considered. Fittings often turned out too small whenever I tried to put the ball into the bearings, however, I solved this problem easily by filing and sanding down the edges to match the dimensions I wanted. These realisations made me rethink the efficiency of iterative design with a 3D printer.
Once all the designing and assembly phases were completed, I had to dabble in some electrical work which required some soldering and wire cutting. This included me soldering wires to motors, soldering wires to LEDs, and shortening cables by re-joining ends of wires together. In the process, I used heat shrinks extensively as it was extremely convenient and easy to use. There were minor difficulties while working with the Arduino and Bluetooth module, as the Arduino Bluetooth terminals did not supply the appropriate voltage to the Bluetooth module. Therefore, I had to reroute and solder extra pieces of wires onto the Bluetooth module to reach the appropriate voltage, resulting in a slightly less than organised aesthetic.
The final design was programmed to do a little dance to demonstrate its ability, and given as a Christmas present. As this is only Ver 0.0, I believe many improvements can be made to the next design such as smaller batteries to minimize chassis size, better designed ball wheels and inclusion of LED lighting patterns (and perhaps even inclusion of high powered lasers to pop balloons?). In the near future, TW BOT Ver 0.0 will receive an update to make it work with nous. Further testing will be needed to fully develop TW BOT, and I am sure there will be many more changes to the designs as well.
As we continue to progress in our work, I continue to anticipate my next task with excitement.
Believing in the potential of brain sensing technology to further our own understanding of the brain. And ultimately, progress our technology into a new frontier with unlimited possibilities from motionless controls, to artificial intelligence, to a fully functioning humble little TW BOT.
