In 2015 a friend and I built a couple prototypes for an idea that I still think about, but haven’t cracked yet.
It was inspired by a couple things: first think at the time I had a pretty uncomfortable mattress i had gotten for free, and second I had seen some work published by the MIT Media Lab’s Tangible Media group on a very cool invention that was being used in all sorts of interesting ways:
A programmable surface consisting of an array of electronic linear actuators. Tremendously cool, and stuck in my mind.
Earlier, soon after graduating college, I had worked for a sail loft based in Hong Kong and got to learn how modern sails are built, fabricated in the shape of the ideal foil for the boat. North Sails pioneered this process in the 90s by creating a surface that could be programmed to a certain curve, upon which kevlar strings were laminated. This set the standard for performance sails.
Basically the idea I had was — wouldn’t it be so comfortable to have a programmable surface for a bed?
We’ve all seen these mattress advertisements:
Of course the problem here is that there’s NO bed that could actually support someone perfectly like this unless it was specifically designed for someone of their proportions and weight distribution, and even then they would have to stay in the same position all night!
Our thinking was that this would be an ideal use case for a programmable surface, because it would find a middle ground between deformation and support, tailored to your actual 3D weight distribution on the bed, over time.
It struck me that this was an ideal application for the concept, but not for the technology — no one (or at least, not me!) would want to sleep on a large noisy array of actuators.
We figured that a successful technology that we could bring to market would have most if not all of these criteria, if we were to adopt a distribution model like some of the hip online mattress companies, and avoid investment in brick-and-mortar:
- < 10,000 USD market price initial target
- Good resolution (between 20×26 and 60×80)
- Easy to Ship
- Non-significant EMF
- Easy to assemble
- Relatively fast actuation (Full range of motion in <5 seconds)
- Accurately measurement of downward pressure at each segment, or voxel
It was clear from this set of criteria that there had to be a new mechanism that required significantly fewer parts, and ideally instead of being an array of discrete actuator systems would find some way to share power generation or actuation between actuators to reduce the number of components by a factor of 2 or more.
This led us to wonder whether you could get away with programming strips of the bed, rather than vertical linear actuators would be enough, since you wouldn’t expect there to be a large height difference between adjacent actuators anyway. We decided to prototype a bed that consisted of strips of nylon webbing, controlled by electric pulleys.
Unfortunately the hard drive that contained the nice renderings and pictures of our prototype appears to have a stuck head, and wont read, after being mailed across the country. I think I can fix it, but in the meantime I’ve sketched a few diagrams to explain a bit clearer:
On the left we have a hypothetical queen-size bed version of the MIT TRANSFORM bed.
On the right, we have our attempt at reducing the complexity for the purpose of queen beds. Each strap is tightened or loosened by a single motor pulley.
We found out two things from this prototype:
- It was very comfortable for a single person
- But untenable for two, because it had a tendency to sag in the middle, and had no way of programming different degrees of support from left to right.
So we added vertical strips, thinking we could solve both problems this way:
Ultimately we realized that the gauge of webbing required to support a person adequately meant that the straps could never be pulled flat, and left a very frustrating dip in the middle. It’s an introductory physics problem, but we didn’t realize how uncomfortable it would be!
Here, taken from Stack Exchange:
So, we were back to the drawing board!
Life took us on different journeys before continuing work on this concept, but I’ve been thinking about trying a hydraulic or pneumatic system out, where each piston has a set of actuated valves that control inlet and outlet to a shared pressure system. Perhaps someday!