Using simple feature extraction to control lights

Projects General Guidance
1

Hi. I’m not sure if I’m using the right words (like “extraction” and “edge”) to ask this question, and if I’ve asked too much in this topic. What is feature extraction called when there is only one edge whose length doesn’t change, and the edge comes from two wearable electronics, on one person? Can it still be the same edge if the edge is moving? I would like for the length not to change; but without a library, I’m worried it might change if the position of one of the wearable electronics changes (because I twist a part of my body one of the wearable electronics is on; or it slides, because what I’m wearing is too loose-fitting). I’d like for the length to be kept the same by using code.

I’d like another edge to come from (one of those two) wearable electronics and a geographical location (that I identify by typing numbers that have to do with at least one distance/angle, or whatever would be best).

In addition to those two edges having a common vertex (that come from a wearable electronics), I’d like for those two edges to be used together (at different times, as they move) to make angles. To make each angle, the vertex each edge has in common will always come from the same one (of the wearable electronics). I would like the angles, but not a screen, and not a virtual reality headset. In other words, the vertices and edges wouldn’t be seen. Then, among other things, the (differences between the) sizes of the angles would be used to control how frequently a light flashes.

Are there libraries for doing all that? What would be the best technology to use? Is saying the wearable electronics stay on a large indoor stage (that signals from outside can’t reach, because the building doesn’t have windows, and is surrounded by taller buildings) descriptive enough? Instead of wearable electronics, would it be better to use something that reflects signals? What is all of the above called? Or, is there a better way? Thank you.

2023: The Year in Review
2

What is an "edge", and how does it come from "two wearable electronics"?

3

Your description makes little sense to me. Can you post pictures depicting what you are writing about.

4

It might be worth drawing something - stick figures are ok.

Then we can relate your words to something tangible, and perhaps there’s an easy way to tie them both together.

5

You're assuming a solution to a problem without telling us what the problem is and then asking if the solution will work.

Start over: what do you want to accomplish?

6

Please see https://docs.google.com/document/d/18Nyu_A3GETyZdo0l9SW81nmH4TGP2bCN/edit?usp=sharing&ouid=102420801851138330130&rtpof=true&sd=true.

7

Thanks, but I see no point in going off site.

8

I got “New users can not upload attachments” and “New users can only put 2 links in a post” messages. Is that what you were talking about?

9

I will not click on the link in post #6 to a google docs site, and neither will most other forum members.

If you want help on this forum, you are expected to take the time to describe the project in sufficient detail, in your posts.

10

I might have answered your question in a post that is too long for some to read, so I’m going to quote myself.

I might be able to use the light to give hints. The light might only give hints when my forearm isn't pointing towards the geographical location as much. The speed with which my forearm points more towards the geographical location would be used (to control the light). After a while, to make the hints harder, what the light does could become more random. So, it would be like a game.

11

An edge is straight, between two vertices, and might include the vertices. If the length of an edge can change then an edge isn’t what I’m looking for. For an example of a length that doesn’t change, in the video at https://youtu.be/-1k3KZCF2NU, I’m assuming the length between computer generated joints (that are next to each other) don’t change (though I might only need two joints). It is okay if the orientation of the edge changes, and the edge shifts. The positions of each of the wearable electronics is supposed to be measured, because they are part of a positioning system. Those two positions are supposed to be used to make a line segment between them. Each line segment might be a different length than the other line segments. I’m guessing the line segment and the edge will be partly collinear; and start at the same point/vertex (for a picture, please see…), end at the same point/vertex, or the shorter one will be in the middle of the longer one. (I could draw pictures of the last two possibilities, if that would be useful.) That is what I want to start with and get, though I’m not sure if it is done that way. Maybe it is done by using more external references (for a picture, please see…), and/or telling a computer two distances.

The two distances would start from different wearable electronics that signals can get to, and both would end at a wearable electronic (on a joint between them) which signals can’t get to (in order to calculate positions where signals can’t reach). I said a segment/edge ended where I said it began, because I was describing different things. I could try to draw a picture of that; but if the joint was bent, it might look too complex because it might involve the intersection of spheres (each of which has a radius equal to one of the two distances mentioned) at more than one point. So, to get all that information, if it is based on something sensible (which I think it is), it might be better for someone to explain using words. Still, I drew a picture in which the joint isn’t bent, so the spheres only intersect at one point. For a picture, please see...

To make an angle, suppose one of the wearable electronics is on my left wrist, the other one (which is used to make the vertex that both edges have in common) is on my left elbow, and the position of the geographical location is used. If I point my left forearm (which is between my left elbow and my left wrist) away from the geographical location then the angle would equal 180 degrees (and the light would flash less frequently), and if I point my left forearm towards the geographical location then the angle would equal about 0 degrees (and the light would flash more frequently). For a picture, please see...

In conclusion, if I can do the above, then I might be able to use the light to give hints. The light might only give hints when my forearm isn't pointing towards the geographical location as much. The speed with which my forearm points more towards the geographical location would be used (to control the light). After a while, to make the hints harder, what the light does could become more random. So, it would be like a game.

Thank you.

12

A magic game.

13

Still not there. As a general rule, I find that if you can't simplify what you're asking into two sentences, then you may not have as complete a grasp of what you need as you think.

What I got from all that is that you need to identify the location of specific points in space so you can determine their relative positions. Would that be an accurate summary?

14

Yes, and I’m hoping experts will know if what is identifying locations will slide around those locations too much, resulting in too much error. Because two of those locations will always be the same distance from each other, I thought maybe the computer should make the points that distance from each other, all the time.

15

All of this is confusing the issue. One thing at a time.

OK, we've determined that you want to identify the location of points in space. Are these points always on someone's body or could they be somewhere else?

16

There would be at least two points that are always on someone’s body. Another point (whose location I want to identify by typing numbers, after I’m done writing the code) would never be on a body, and could be “far away” from those two points. By “far away”, I mean anywhere in the world.

17

OK, now we're getting somewhere in terms of what you need.
What is the required accuracy of the position measurement?

18

Thank you. I’m hoping someone can answer that question for me, if they knew the output that I want. Is that okay?

19

Should that be in a new topic?

20

No, keep all related stuff in one thread.