December 8, 2019 · project

Reflections on Prototyping

Partially constructed MVP prototype

Last week I started prototyping for my final project. The MVP prototype I created consisted of a series of capacitive touch sensors, lined up and sandwiched in between two pieces of fabric for insulation. I was quite pleased with the success I had at being able to register a stroke across the sensors, although the way I coded the detection could definitely use some work. No better was this demonstrated than with my attempts at data visualisation using the Serial bus and Processing. Initially I tried to pass my data through as it was in my original prototyping code.

Processing didn't seem to like having that much data thrown at it at once, though, and in the end I rewrote the code so that the Serial connection was dealing with less data. The trade-off was a slower detection of touches on the sensors, though, as  I had to significantly slow down the rate at which the sensors checked for changes in their conditions.

These two experiments in combination have lead me to the conclusion that capacitive touch sensors alone may not be the best fit for my project. They can be quite unstable, producing wildly varying data depending on such small things as whether or not the user of the device is wearing shoes, or leaning against a radiator (both things that happened during testing). A potential replacement for my capacitive touch sensor design may come in the form of the Piezoresistive Sensor Matrix, or "Pressure Matrix" [1]. Essentially a grid of wires that meet at varying points, when cross sections of these wires meet, we can sense varying levels of pressure across a grid.

Piezoresistive Sensor Matrix by Kobakant [5]

Perhaps somewhat ironically, the pressure sensing capability is not what makes the matrix potentially so useful to my project, but more the grid layout, and the ability to pinpoint where on the grid is being pressed upon the hardest. If I can track this across a surface, I potentially have a system that could detect stroking and petting in multiple directions, rather than just the two of my prototype design.

Visualisation of my potential use of the Pressure Matrix

There are challenges to using the pressure matrix in my project, however. The electrical engineering/mathematics behind the design is something I'm struggling to get my head around, so some more research in that regard will be required. Another challenge is that the design uses fabrics that are difficult to acquire, so some compromises may have to be made on the quality of the materials used.

The pressure sensors would be placed either side of a central pouch, containing the main components of my project. At the moment I'm envisioning a sandwich of fabric, sensors, and components. I've settled on a round, spherical design for the shape of my device. I thought about the shape of my hands, and the way they cup around mugs when I'm drinking coffee, or holding a small animal like a mouse or a shrew.

A round shape feels very friendly and familiar, and that's the way I want to take my design. The downside is, of course, that this will challenge my fabrication skills. Dealing with flat felt sheets in my prototyping is one thing, dealing with stuffing, and a three-dimensional housing will be it's own challenge.

Diagram of the "component sandwich"

One aspect of my project that my prototyping has not yet approached is the response that the device will have after being stroked/petted. As mentioned in my project proposal, I'd like it to purr at the user, to illicit a feeling of safety and calmness. If the device responds the way a happy pet would, it should be a calming experience. My intention is to use vibration motors to produce a sort of cat purring. Getting the feeling of this purr correct is important. It has to feel natural.

There are some rather lovely studies on the purring patterns of both domestic and wild cats [2][3]. They have mostly been carried out by I intend to study the results of these surveys and look at both the rhythm and intensity of the purring from different cats. I'm expecting that getting a consistent yet not robotic pattern is going to be tricky, and that a lot of tinkering with the pattern will be required.

Lastly, I'm yet to pick a material for the main body of my device. I know I want it to be soft, and I know it needs to feel natural to want to stroke it. A popular fabric for home-made soft toys is Minky [4], a soft polyester fleece-like material. It can be quite pricey and apparently leaves dust everywhere. It does come in lots of fun colours, though!

Some home-made Minky friends (they were gifts).

I had a lot of fun with my prototyping. Although it still feels basic, it's progress towards the end product. I've lots to do going forward, but I'm very excited!


[1] Kobakant Blog - Pressure Matrix
[2] - Devoted to Felid Purring
[3] A Comparitive Acoustic Analysis of Purring in Four Cats (Schötz & Eklund 2011)
[4] Online Fabric Store - Minky Product Guide
[5] Peizoresistive Sensor Matrix example