Hardware components | ||||||
 |
| × | 1 | |||
 |
| × | 3 | |||
 |
| × | 1 | |||
| × | 1 | ||||
| × | 1 | ||||
 |
| × | 1 | |||
Software apps and online services | ||||||
| ||||||
Hand tools and fabrication machines | ||||||
 |
| |||||
 |
| |||||
My project asks: can technology open up new perceptual realities without becoming the thing we rely on to access them? And if so, can a wearable device assist in the acquisition of human echolocation as a permanent, device-independent skill?To answer that, I built the ECHO V1, a wearable device that works like training wheels for echolocation, designed from the start to be outgrown.
Fabrication processThe fabrication process took around 6 weeks in total. Instead of waiting to cram in everything at the end, I started working before the MDEF Fabrication course started. That allowed me to test a prototype beforehand, which I did not like. Luckily, I had time to pivot and iterate.
The first iteration is shown above (just the breadboard, not the integrated design of it. Instead of using haptic motors to give feedback from the ultrasonic sensors, I tried to use audio as feedback. It worked well in terms of understanding spatial distance, but it didn't make sense for my research question because the user wasn't able to hear their own echo, thus there could never be a coupling of echo and feedback necessary for learning the skill without technologically mediated assistance.
I decided to pivot to haptic motors. Since I had experience using them to convey information in my previous cognitive orgy projects, it felt like a smooth continuation of what I had already been working on.
Here is the very first prototype using cardboard as the wearable material. Look at all those cables, it was a total mess. I spent over a week on cable management, which was incredibly tedious, but also super satisfying by the end.
Integrated DesignOne of the most rewarding parts of this project was the lack of AI's ability to help me. I have no ethical qualms using AI to help guide my work in disciplines I don't know much about (like wearables and integrated hardware). I tried to have AI generate a schematic for me, but it was completely unable to produce anything useful. Paradoxically, this ended up being a huge help because I had to sketch everything by hand, which helped me understand every step of each design choice. I had A LOT of issues along the way, and feeling like I knew how to fix/troubleshoot them without outside help (I'm looking at you, Claude) was super empowering.
Above is a schematic of how I accomplished my cable management for the motors, sensors, and mic.
Above are the hand drawn wiring schematics for the flexible breadboard. I had to do everything backwards since I was using the back of it. I ended up giving each row/column a letter so I could refer to spaces easier.
Above is the finished, soldered circuit. It took many hours of soldering. The process was mostly meditative, interspersed with periods of intense frustration. Soldering and juggling are the only two activities I engage in that viscerally remind me of our evolutionary appendage limits, and make me wish I had at least two other arms and hands. Maybe if I keep on doing them my descendants will evolve with those adaptations.
STORY - Design Process (Describes design strategies, steps, problems and how the team solved them - iterative thinking)As I mentioned earlier, the design process took around 6 weeks from start to finish. I learned from previous projects that when working with electronics, it is much more advantageous to work a little bit each day than try to do it all in a sprint at the end. I started the course with a rough prototype of the cardboard with the sensors embedded. For the first two weeks I worked on cable management and getting all the components set up on their own. Shyam gave me the idea to use a medical face shield as the frame for the device, since they're sturdy, have plenty of space to mount things, and can be adjusted to fit different sized heads. After I had all the components wired individually, I started by embedding the sensors into the plastic face shield. I made space in the plastic by outlining the sensors and then burning a hole with the old soldering iron in the MDEF room. After those were placed I had to figure out how to mount the microphone. I was going to 3D print a microphone holder, but then I found some copper wire. I wrapped it around an edge of the helmet, bent it in the shape that I liked, and found a wire coverer to keep the microphone cable in place. It worked perfectly.Once those were sorted I had to glue the motors to the headband. That was an arduous process. I decided to take a shortcut and salvage motors from a previous project. I didn't like how it ended up because the motors wouldn't lay flat. Looking for inspiration, I saw how Erandi wired the motors for her project, and decided to copy that layout. It added at least 5 more hours of work, but it was worth it. After all the components were in place I soldered the wires on the flexible matrix (as shown above), and used AI (Claude) to generate code that would run tests to make sure each component worked properly. You can access the code for these tests in the Attachments section of this project.Once everything was up and running I attached the matrix and a portable power bank to the headband with zip ties (I've come to the realization that zip ties are one of the most under-appreciated tools available in the FabLab). After everything was integrated I used Claude to generate the code for the navigation, using the previous test codes as a foundation. After some tweaking and user-testing with other people, I settled on code that worked right.
Everything worked perfectly until the day before the event, when the microphone abruptly stopped working. I tried everything to fix the problem, but unfortunately all diagnoses led to one conclusion: the microphone was broken. Luckily, Dani had given me two when I was starting the project, so I went back to the FabLab and soldered on a new mic. Lesson learned: ALWAYS grab a backup component.
References- Nagel, T. (1974)— "What Is It Like to Be a Bat?" — The Philosophical Review, 83(4), 435–450.
- Teng, S. et al. (2015)— A device for human ultrasonic echolocation — IEEE TBME
- Thaler, L. et al. (2019) — Human Click-Based Echolocation of Distance: Superfine Acuity and Dynamic Clicking Behaviour — JARO
- Castillo-Serrano et al. (2021) — Increased emission intensity can compensate for the presence of noise in human click-based echolocation — Scientific Reports
- García-Lázaro & Teng (2025) — Novel Stimuli to Benchmark and Train Echolocation Skills
- Ekkel et al. (2016) — Learning to echolocate in sighted people — Experimental Brain Research
- Human Echolocation — Wikipedia
- SEP Sonic Eyewear Project
The ECHO V1 worked as a navigation tool. I was able to use it to navigate through spaces, as was everyone who used it during MDEFEST (to varying degrees of comfort). However, that is not enough. The ECHO V1 is not meant to be a navigational tool, it is meant to be an echolocation training tool. To understand if it can accomplish what it was built to do, I have to test it for weeks at a time, with various subjects, following a standardized method, and then test again without the device present. The bulk of future development starts there.Aside from testing, I would also like to experiment with camera based distance sensing instead of the ultrasonic time of flight sensors. I think it would give a more detailed report of the area around the user, and wouldn't be limited to objects at eye level. However, that would open up a larger issue of how to translate that into haptic feedback. So I think the first thing I would like to further develop is the training of the existing device before I try to iterate further on a new one.
Comments