I decided to experiment with the smart multi’s today. I had two extras after my move. I had the crazy idea that I could strap them to the chairs in my home office and use them as presence detectors. It works. But in looking at the data, I think I could do more. Just wanted to put this out to the collective and get some opinions.
My first experiment was trying to detect taps on the arm rest. This seemed to work OK, but will require a new SmartApp.
Second experiment was trying to detect if I am leaning back too much or not maintaining the correct posture.
I am trying to correct my chicken head posture that I have created by spending two much time hunched over a keyboard. In theory this could even be integrated with some sort of haptic feed back mechanism. A simple motor for vibration feedback embedded in the chair that fires after a given period of time. Or make your jawbone pulse a notification. Too many possibilities.
I was discussing the concept of having a chair sensor with @tgauchat just a few days back. The problem is that we don’t have an easy way to do this using a sensor dedicated to detect occupancy of a chair. While accelerator based approach may be a plausible approach, I dont think it will help me know the current occupancy state of a chair.
The use cases I was thinking of are on the same lines of what people have been suggesting in the thread. But it can also be potentially expanded to small businesses like local pubs, restaurants etc where the state of a bar stool or chair can be made available on their website or just help them run some useful analytics - Which will help their customers plan better and help the businesses meet their customer needs better.
I was thinking of a cheaper and effective way of doing this but it may call for new hardware development. Ideally, it should be something that is easy to place on any kind of chair (wood, sofa, recliner, wheel chair etc) and secondly it should be cheap - May be something less than $25 per sensor, so that people can afford it for all their chairs and Thirdly, it should be transparent to the user.
Some stores implemented a sensor, for every time someone plays with a product, how long they pull to play with the product, heat maps of total walking areas. We knew by color, we had play + retail interfaces that also track patterns…
Music + volume + type of genre would play by daytime + amount of people in store + other data.
It was AWESOME. best way to predict what someone does in a store.
That’s a first step of “Smart” Stores… which, is an approach for SmartHomes too.
There are pressure sensors that can be wired to a zwave contact sensor. When you sit down the contact sensor “closes” and thus the seat is occupied. I’m currently using this setup for my front door mat. When someone stands on it, it turns on the patio light. If it stays “stood” on it sends me a txt that I have a package waiting.
Lots of hospitals already use pressure mats to detect occupancy in chairs and beds. These are relatively cheap and easy, although you’d have to be buying in large quantity to get under $50. But you can make your own for a few bucks, and then the only real cost is the cost of the antenna. You can definitely put the pressure sensor under a regular cushion and leave the antenna piece attached to the chair frame…
Both chairs in my office are pneumatic, and as such, they give a little when one sits in them. They also rise back up when one leaves the chair. I am looking at the coordinate data from the ST multi and there is a very predictable amount of movement along the y axis. I am going to throw together a smart app that treats the chair as a contact sensor or button. I will allow for a selectable range of movement so others can use it as well. I think this will be a very good indicator of “chair presence”. This should allow for a whole host of chair related activities to be used in other smart apps.
Pressure mats work well in hospitals, beds and wheel chair applications. But they are bulky and i am sure nobody would want a pressure mat on their fancy dining chairs
I propose a sensor that is probably as big as a thumb and you can place it under the chair right at the edge. Implementation can be done in many different ways , but as long as the sensor detects proximity or change in capacitance ,Its state will be “occupied”… Otherwise it will be “empty”… Any other implementation ideas?
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
16
There’s work on detecting room occupancy, including accuracy to number of occupants in a room, using measurements of radio wave (RF) interference. Some of these techniques even work through walls.
I’m tempted to think that some sort of high-frequency or RF pulse would definitely respond back to the sensor with a distinctly different pattern when a chair is unoccupied vs. occupied. Certainly worth an experience … if you can decide on what type of emitter / receiver hardware is worth testing to start.
This might go outside the boundary of what you’re trying to do, but I thought about using an Arduino with ThingShield and a simple 5v motion sensor to create some sort of presence sensor just for a single room or a desk area. You could write the code so that it reports motion events based on just a few seconds of motion/no motion, but wait for up to 5 minutes or so without motion to mark the room as not present.
Room occupancy based on a single motion sensor tends not to work in practice. Zone detection (multiple sensors) is good for turning lights on, but can still be tricky for triggering lights off.
The most accurate for room occupancy is probably a double trip beam where you can figure out which one tripped first and so distinguish coming from going.
Discussion of all of these here:
For a chair sensor, it depends on the exact use case. I disagree that pressure mats have to be bulky–they can be very thin and small, just a strip under the cushion. Coaster size, even. Often used in amateur robotics. Easy to make as well.
But certainly there are many use cases where an accelerometer or vibration sensor might work as well or better.