Why are there so few open-source Indoor Positioning System (IPS) implementations

Using Arduino Programming Questions
1

I've been doing research for the past few days and there seem to be hundreds of academic papers covering IPS over the following technologies:

Visible Light Communication (VLC)
Computer Vision
Sound
Magnetic Fields
Dead Reckoning
Ultra-Wideband (UWB)
Wi-Fi
Bluetooth Low Energy (BLE)
Radio Frequency Identification (RFID) and Near Field Communication (NFC)

These academic papers contain the math and algorithms the researchers used, but they don't show the code that they used and there aren't many open-source projects on GitHub either.

What gives? It seems that everyone here would want a robot to be able to drive around their house, so there is much interest but not many implementations.

2

What gives?

It’s hard and not well suited to implementation on small processors. A lot of the academic work is subject to IP restrictions by the people sponsoring the research.

Nothing to stop you from reading those papers and trying it yourself though.

3

Invention doesn't work by dreaming up something that we would like, and then producing it. It happens when existing technologies ripen and mature to an extent where they can be combined into something else. Visionaries only have visions until that stage is reached. But many people like to get a head start and also there is pressure to publish papers if you are an academic. So you will see far more speculative research in areas like this, than ready to roll solutions. It's because we are not yet in a position where the required technologies are facile and cheap enough to do it.

1 Like
4

It seems that everyone here would want a robot to be able to drive around their house, so there is much interest but not many implementations.

Absolutely, but it is VERY difficult to implement.

If you've checked the details and prices for the MarvelMind and Pozyx systems, which have been around for years now, you know that they are expensive, and yet have severe limitations.

They work in only one room, and (for example) the MarvelMind locator doesn't work unless at least two beacons have direct line of sight to the rover. So forget having furniture, or people moving around in the room.

It will probably be years before a cheap, reliable indoor positioning system comes out, and you can safely bet that proprietary technology will be involved.

5

aarg:
Invention doesn't work by dreaming up something that we would like, and then producing it. It happens when existing technologies ripen and mature to an extent where they can be combined into something else. Visionaries only have visions until that stage is reached. But many people like to get a head start and also there is pressure to publish papers if you are an academic. So you will see far more speculative research in areas like this, than ready to roll solutions. It's because we are not yet in a position where the required technologies are facile and cheap enough to do it.

But the papers I'm referring to aren't theoretical, the authors have implemented something and tested and written about it. And there seems to be hundreds of these IPS papers. But they don't make their code and libraries available, but they're also not selling a product. Maybe they're worried if people tested it, it wouldn't work?

6

But they don't make their code and libraries available

Nonsense. You can't possibly know that, after "a few days of research". Write to the authors and ask for the code, or better, ask if you can collaborate with them on further development.

they're also not selling a product

Of course not. It costs a lot of money to develop a product. The options are to have a very rich uncle who is willing to fund your team for a few years of work with zero return, or to start a company and convince investors to fund your team for a few years of work with zero return, or Kickstarter. Make sure you have some very good ideas before you head out there!

7

faatma:
What gives? It seems that everyone here would want a robot to be able to drive around their house, so there is much interest but not many implementations.

Do you think its just possible that its a lot harder to implement such systems, in an economical way, than you think ?

If it was cheap and easy to do, you would be able to buy the systems on eBay and there would be gazillions of 'Instructables' describing how to do it.

Perhaps develop the product yourself, the market would likley be huge so you could make millions.

8

What gives? It seems that everyone here would want a robot to be able to drive around their house, so there is much interest but not many implementations.

Do more research. I find several projects with IBeacon / BLE based Arduino projects, including the usage of an ESP32. This is just a first example iBeacon orten mit ESP32 Bluetooth RSSI – ESP32 Arduino Projekte

9

Sometimes it needs the cheekyness of a non-professional with a lack of experience" to just try something new.
And sometimes they make the breakthrough.

One example for this are blue LEDs. Today Nichia is one of the biggest LED-manufactures in the world.
Well **Shuji Nakamura **started as electronic engineer at Nichia on a quite knowledge-level about LEDs
when nichia was only producing fluorescent tubes,
and developed some breakthrough-steps for producing blue LEDs.

Another example is Extinguishing burning oil wells. Some Czech people mounted a big jetfighter-engine on a tank for
extinguishing burning oil wells. Your first reaction might be "whaat??! the biggest hot air dryer you can buy on earth shall Extinguishing burning oil wells??!!" Yes - if you put a lot of water that evaporates immediately into the stream it is very effictive in that. "Red Adair" an us-american very experienced burning oil wells-Extinguishier would have never come up with this idea. Because of his experience.

There is an effect I call the arrogant ignorance of the most experienced experts.
So a lot of different people are good to push development forward some hestistaters and some cheeky experimenters.

Just some ideas:
How about using different ultra-sonic frequencies to have more pulses available at the same time without disturbing each other
How about using a lot of ToF (Time of flight) sensors

Jery Elseworth worked on a project castAR which used triangulation for determing position.
But I don't know if this system needs free sight.

castAR was abandoned by the investors. But there seems to be a new similar project tilt five

best regards Stefan

10

Actually, Boeing did this inside their 747 final assembly plant about 25 years ago. They reported they could position wings, etc within 1-2 mm accuracy using ceiling mounted devices. They reported the system was based on the GPS satellite system.

Paul

11

Sometimes it needs the cheekyness of a non-professional with a lack of experience" to just try something new.

Many people will read that as unskilled people. In order to make a breakthrough you do need a high degree of skill, just not necessarily in the field you are going to tackle.

12

jremington:
Absolutely, but it is VERY difficult to implement.

If you've checked the details and prices for the MarvelMind and Pozyx systems, which have been around for years now, you know that they are expensive, and yet have severe limitations.

They work in only one room, and (for example) the MarvelMind locator doesn't work unless at least two beacons have direct line of sight to the rover. So forget having furniture, or people moving around in the room.

It will probably be years before a cheap, reliable indoor positioning system comes out, and you can safely bet that proprietary technology will be involved.

I have to disagree. Although commercial robots are expensive DIY ones don't have to be and also don't have to be so constrained.

Over 35 years ago i had a neighbour that was a medical assistant, he was also a Commodore 64 nut. This computer was soooooo fast, 1MHz from memory with 64K of memory. With some interesting mods he turned it into a C256 by paging 3 extra banks or memory. All this with no formal electronics training.

He built a robot car that had U/S sensors, IR sensors, both used for distance measurement. He used small RF modules to communicate with a baud rate of 1200 I think.

The car could drive around the house, room to room and never hit an object. You could move furniture and at could avoid it. Part of the communication was mapping of its path, this was done by multiple speed sensors on its wheels. It could calculate the difference in speed of each wheel as it changed direction. After 3 or 4 laps around the house it would have an almost accurate map of the house.

When I saw what he had done I was totally dumb founded. It was beyond my comprehension and I was an electronics technician.

His other little project was a suspended ping pong ball. He coated it in iron filings and it would suspend in the contraption. It had an electromagnet at the top and U/S sensor at the bottom. If the ball got closer to the sensor it would increase the electromagnet current and suck the ball up again. 2 very amazing designs on the most basic (in todays terms) computer. Having said that the C64 was so way ahead of anything else available at the time for the home market.

13