As part of the Create[Space] Artist-in-Residence program at Wyndham Tech School, I've designed a project that explores a hopeful and optimistic vision of technology through a Solarpunk lens. The installation centres around BTN, short for Bioremediation Terrain Navigator, a small four legged companion robot designed to act as a protector and healer of the environment.
This piece and its companion media will be installed permanently at Wyndham Tech School in the hopes that it stands as a physical reflection point for students reminding them to design and build technology that works in synergy with nature rather than against it.
Set in a potential future where society has rejected corporate exploitation in favour of building tech for the greater good, BTN serves a vital ecological role. From soil restoration, navigating difficult terrain to detect ground impurities and performing bioremediation through injection of organic mycelium (mushroom spores) enzymes directly into the soil to break down pollutants and revive local ecosystems.
Using the tools and 3D printers at Wyndham Tech School, this project page aims to document the entire process of bringing BTN from concept to final deliverable. The process will involve large scale 3D printing, assembly, filling, sanding, painting, and texturing to achieve a real world and outdoor worn aesthetic for the final piece.
Before jumping into 3D modelling the first step was defining BTN’s visual silhouette. Since BTN’s main role is injecting mycelium spores to restore soil I decided on shaping his head in the form of a mushroom cap.
To give BTN a friendly expressive personality, the head structure is designed around a single, large optic LCD housing imitating an eye scanning its surroundings. The integrated screen will allow BTN to display animated eye movements and blinking in intervals giving life and character to the otherwise non-functioning sculpture.
With the form, proportions and key feature housings finalised, I selected a colour palette I felt would convey a calm and friendly tone but would also stand out from the green filled surrounding of the forest.
From here, the concept was ready to move into Blender to begin hard-surface modelling and prepare the individual parts for 3D printing.
Using the blueprint sketch as my reference, I began modeling BTN’s core chassis in Blender, opting to tackle the legs first as I needed to nail down a stable base to structurally support the oversized head. Establishing the lower limbs early gave me an immediate anchor for the robot's overall scale, stance, and personality. It allowed me to balance the visual weight of the design before diving into the complex curvature of the upper body, ensuring the character felt grounded, functional, and ready to stand on its own two feet.
Once the legs were finalised, I moved on to the head section which proved far more challenging than anticipated. Translating the smooth and organic curvature while cutting in clean mechanical details required several iterations. I even dove back into a few hard-surface tutorials to brush up on my topology technique to pull off the curved panel lines without distorting the mesh.
Once the legs were finalised, I moved on to the head section which proved far more challenging than anticipated. Translating the smooth and organic curvature while cutting in clean mechanical details required several iterations. I even dove back into a few hard-surface tutorials to brush up on my topology technique to pull off the curved panel lines without distorting the mesh.
Lastly, I designed a dedicated display stand to securely support BTN's weight while anchoring the character in its environment. With both the robot and stand model finalised, the digital build is complete and ready to be broken up into smaller pieces to be printed in large format on the printers. I also appled textures onto the digital verson of the model and rendered out some shots as a digital representation of the final deliverable.
Before sending BTN to the 3D printers, the entire model had to be converted into individual modular objects of print ready parts. This meant breaking up every piece of the model and adding alignment pegs for the pivot points for assembly.
One of the main engineering challenges was making sure the leg joints didn't shift once assembled. To fix this problem I designed a custom star peg that fits inside the joint connection. It’s not meant to be a movable joint after the fact but instead serves as an indexing tool during the first assembly, letting me set the exact angle I wanted for the leg pose in the initial assembly stage.
I chose this approach because I didn't want to rely solely on glue to hold the weight and pose of the legs in place. From a structural point of view, having a mechanical interlocking system gave the frame significantly more strength and ensured the posture locked in place permanently with zero risk of slipping.
To avoid wasting material or discovering issues midway through the large scale print, I printed a complete 30% scale prototype as a test. Printing the small prototype ended up being worthwhile because I discovered a few flaws in the design including an issue with the star pegs that needed to be adjusted and some scaling issues with the fit on some parts.
Now that all the issues have been ironed out, I'm moving onto printing the large parts using Wyndham Tech 3D printers, specificially the Bambu Lab H2D. This is a high-end FDM printer and I chose it specifically for its large print area, allowing me to print larger sections of the body without making too many cuts.
While BTN is printing, myself and Taylor, a staff member at Wyndham Tech, have embarked somewhat of a side quest to get the 3D model working in Unreal Engine for potential use in their gaming education program.
Due to only having a rudimentary rig applied to the current BTN, I set out to give BTN a fully functional armature rig that would allow him to move like it should in the real world but also have the ability to import the model into UE5 from Blender without issue.
After spending some time working on a new rig, I figured out a much more efficient way of setting up the connections from the body to the legs and included a root bone for overall control. This final version worked as intended when importing into UE5 and over the next week will work with Taylor to get procedural walking working within Unreal Engine 5.
Once the rig was finished it was time to get some simple walk cycles going and bring BTN to life. To figure out how a cute four-legged robot should actually step, I ended up watching a bunch of slow-motion videos of dogs walking to understand how they move. Going through this step helped me get the leg sequencing down so I could tackle how this kind of robot would go about moving around.
Firstly, I got the four-leg timing down, then added a bit of a head bobble so the animation doesn't look totally stiff when it steps, sort of like suspension compensation. To keep it from looking like a rigid loop, I added in some random noise to give each bone a little bit of organic shake and sway which makes it come together and feel more alive.
Another potential idea from Taylor was making a functional moving version of the smaller BTN robot using some of the spare motors at Wyndham Tech, but this project will come much later on after my time at Create[SPACE] and would be something I would continue to work on after the program.
This moment for me has been a great reflection point as I feel like this project has now taken on a life of its own and opened more doors for me to explore even after this program has concluded. Making meaningful partnerships within Wyndham Tech has been something I am extremely grateful for and wouldn't have happened without the Create[SPACE] opportunity.
Production is underway! I started printing the final components at Wyndham Tech using their large format Bambu Lab H2D printers. Being able to run these on bigger machines meant I could print large sections of the head and body in fewer pieces, meaning overall it would be a sturdier build and get done quicker.
This part is always the most fun, slowly watching the parts come together and build the robot piece by piece, while my personal 3D printer has printed the smaller leg parts and the lower section of the body core.
While printing for the head is happening, I begin sanding and applying the first layer of filler primer to all the leg pieces to get this part of the process started. I've also ordered all the paint and some had arrived so I'll be able to begin painting the metalic colours in soon.
Having painted the lower body parts I then weathered them. I either used a liquid mask rubbed off with steel wool to create a chipped paint effect or added scratches with the steel wool directly to the pieces.
Having completed the body and waiting for the head to finish printing after a few failed attempts, I’m shifting my focus to creating a small advert for BTN. This advert will showcase all the parts coming together and promote it as if it’s being showcased in a future world. Below is a rough preview of the first three shots from this side project.
While at Wyndham Tech, I've begin working on getting the electronics sorted for inside the head. Taylor has been a massive help in getting me started and helping me understand the basics of electronics, also just bouncing ideas off and getting a rough idea of how everything could work together.
With just two and a bit weeks to go, this last stretch has by far been the most concentrated learning experience on the project. From diving head first into the Arduino Uno microcontroller ecosystem, to learning about stepper motors, sensors, wiring, power management and even blowing up my first capacitor, learning that yes, a capacitor does indeed have a positive and negative.
There have also been a few huge moments since my last post, the first was getting the stepper motor to simply rotate. Sounds easy but stepper motors are their own specialised type of motor and require a separate controller to translate both power and input data into very precise movement. After a few days of tinkering and many YouTube tutorials, it was quite exciting when I got it to rotate for the first time without it glitching out.
Next, getting the motor to directly control head movement was the next objective. After talking with Wyndham Tech staff, one of the staff suggested an extremely simple peg solution (shown in the pictures below) to directly control the head, rather than using gears or belts. This was perfect for me because it was a basic but quickly achievable solution to the problem.
Using a hot glue gun and some spare pieces of wood, I created an extremely primitive peg setup and did an initial test with the motor inside the head. To my surprise, it worked! I had to dial in some settings but the head did a very basic back and forth sweeping motion, but it could handle the weight of the head and moved in a way that gave me confidence to move forward with this method.
This was also the first time BTN had any signs of life, which was an exciting boost at this stage of the project as I've never actually used motors or microcontrollers in my projects before, so to get some early signs of it finally working was also a relief.
Now that the head movement was sorted it was onto figuring out how to move the ears. I wanted the ears to move because it felt like that would be the best way to inject some personality into the robot and would pair nicely with the eye animation.
Getting the ears figured out ended up actually being easier than I expected thanks to all that I picked up from working with the stepper motor, not being so worried to just assemble the roughest version of something to experiment..
Up until this point, I had been overly cautious as I didn't know if what I was doing would even function, but thanks to the success of the stepper motor working, I felt like I was starting to get a grasp on the idea behind the general process.
Figuring out the ear rotation ended up being fairly straightforward. I wired up two micro servos that each talk to the Arduino Uno and are connected via a drilled hole in each side of the head. I then glued the end of the servo to the ear and housed the servo on the head interior, testing each individual ear, making sure they are able to make full rotation without getting stuck. Both worked out on the first try, which was incredibly lucky.
I then coded a very small 'animation' via Arduino IDE that essentially does a small wiggle at a random interval, adding some subtle emotion to the robots overall personality, and allowing the viewer to project their own emotions onto the meaning behind the ear wiggles.
Another feature I added in this stage was a sensor that would activate BTN when someone approaches and would also made BTN go idle when there is nobody around. I wanted this feature mainly to give the viewer a sense that BTN was reacting to their presence, and would start emoting when someone approaches. It also save power and allows the motors to rest while there are no onlookers.
This again felt like it came together extremely quickly, and I got a working prototype within a day of tinkering around. It works using a sonar sensor that picks up changes in object distance, so when there is an object close by, it will activate the motor sequence, and when the object distance is far away, it will deactivate the motor sequence and go into idle mode.
This sensor will be placed inside the front end of the base, and I hope to make it as discrete as possible by hiding it within the base decoration. This feature ties the illusion together that BTN is a functioning robot that reacts to you.
One of the final pieces for BTN's appearance to be complete is getting the eye component installed. Early on I knew that the defining feature of this robot character would be an LED matrix display that I could input a custom animation featuring a digital eye that could blink and look around.
I also knew fairly early on that I wanted a honeycomb lens over the matrix panel acting as a diffusion layer that would disperse the light and hopefully create some depth to the overall eye component. This lens effect was also chosen because I feel like it would also just be an interesting effect rather than going for a traditional LCD panel and doing effects as a digital overlay.
First component added in was the honeycomb lens, fitted in using a custom 3D printed housing that secured the lens comfortably to the opening on the head piece, making sure the lens stays put and doesn't fall out if it's poked or knocked.
When i reassembled with the lens in place, I had a real wave of excitement wash over me seeing BTN in what felt like its complete physical form for the first time since starting the project. The lens ended up helping to sell the overall aesthetic and gave it a legitimate sci-fi vibe which was exactly what I was hoping for.
Next up it was time to add the LED matrix to the cutom housing. What took a whole month to arrive, the part finally arrived and was time to get the glued in and start testing how the matrix would actually look paired with the honeycomb lens.
With everything installed it was finally time to test the LED matrix paired with the honeycomb lens, using a temporary eye animation I quickly threw together. All in all, I'm extremely happy with how the eye looks and this was a pretty cool milestone seeing BTN come to life and. In the end the eye component really felt like it pulled the entire design of the robot together and the lens gives off a really wild 3D effect in person.
With the last week of the Create[SPACE] program having arrived, it was time to add any last minute touches to the project before finishing up and delivering the final piece to Wyndham Tech.
So far the only things left on the checklist to complete was finishing up the base that BTN stands on, and wiring the power cables for BTN's eye and motors through the body.
The base has been something I've been heavily procrastinating since becoming enveloped in the robotics side of the project, so making time this week to figure out how to decorate the top of the base was top priority.
The first step involved sourcing a specific clover style grass, which proved much harder to find than I initially realised, but after a day or so of searching I finally found something I was happy with at a local variety store.
Next was texturing the assembled base with modelling putty by repeatedly dabbing a foam brush, coated in the putty, onto the surface to create a rough textured look.
Once that was dry, it was onto painting using a brown colour for the dirt layer and a green mixed with some black for the mossy grass layer. The paint didn't have to look too realistic as its purpose is to give a convincing layered effect, like the clovers are sprouting out of real dirt underneath.
The final step was hot gluing hundreds of individual glover bunches into the base untill the entire top surface was covered. After an hour or so of gluing, I finally finished assembling and could step back and take in the work.
The base turned out much better than expected and can't wait to unite BTN with the finished base to finally see how the piece comes together as a whole in this final week.