Precise (+-2cm) indoor navigation system for robots

Hello guys,

We have a precise (±2cm) indoor navigation system (Indoor “GPS”) that is designed to provide real-time location coordinates for autonomous robots, vehicles.

Demo how it works:

Basics: You have stationary beacons (not Bluetooth or WiFi) every 10-50 meters. And you trace a mobile beacon on your robot with ±2cm precision. You get the coordinates either directly from the beacon on your robot via USB or UART or SPI or I2C; or you can get data from the central controller - depending on the needs of your system.
Detailed description of the protocol is available.

More information is on [http://www.marvelmind.com/
I will be happy to provide you with more information, if needed.

The most important is that the system is readily available and you can get it within 5-7 days (shipping time) :slight_smile:

Kind regards,
Maxim

email: info@marvelmind.com](http://www.marvelmind.com/)
marvelmind_indoor_navigation_system_ENG_19_details.jpg
_IMG_7658_800.jpg
_IMG_8554_800.jpg

I rather like this actually. It’s not a bad alternative for something like the RPLidar which requires a bit more processing power (and a hundred extra dollars).

I’ll probably still be sticking to kinect and lidar SLAM for my own projects since I need them to work in any environment without configuration, but this would be rather interesting to incorporate into an offseason VEXU competition, maybe set up a system to send the location of ALL robots on the field and see if its possible to play a fully autonomous game.

edit: What is the range on these?

I like RPLidar quite much, actually! Let me check more. But first impression - cool!

Answering your questions:

  • Some guys who have bought our system, they do have two LIDARs and Kinect on their robot already installed. Nevertheless, in case of being crowded, the robot is lost. And it needs some fundamental references of location to realize where it is. To reset. Sort of absolute GPS coordinates. That is what we provide with our system.

Certainly, you need to sense the environment with something (LIDARs, Kinect, basic ultrasonic, infrared, optical recognition, cameras).
But your coordinates - it a different story. You don’t need to employ complex and less reliable and less precise systems. This we solve pretty well now.

Beacons can be up to 50 meters (160 feet) apart.
With a dozen of beacons you can cover very large indoor space.

The only fundamental requirement is the same as with GPS. At any given point in order to have exact coordinates, your robot must see at least three stationary beacons (“satellites in GPS”).
With partial coordinates it can be less than three. With this SW we are still experimenting.

If there is a sudden ultrasonic shadow or something, the system automatically switches to other available beacons. Thus, it is recommended that more than 3 beacons are visible to the robot. In such mode, passing people, etc. do not create loss of coordinates. Besides, we recommend to put beacons on the ceilings to reduce the probability of shadows.

Actually, SLAM with kinects and LIDARs are actually really accurate, especially after loop closure. I’ve gotten localization down to 2cm (measured by starting the kinect in a marked location, mapping, returning to said location, and measuring accumulated error)

I imagine you could get much better results with an RPLIDAR feeding data in, and even better if you had a nice HOKUYO.

However, 50m is rather impressive. I coult see this even used in FRC if enough people got behind it.

Alright. now one more question for you fine folk. How many mobile beacons could you realistically track in one space?

Please, understand me correctly. I like LIDARs very much :slight_smile:
And we aare not even trying to replace them. They serve their purpose. I simply highlighted that the guys with the robot already equipped with 2xLIDARs and 1xKinect decided to install our system on top, because they still see the benefits to their existing and pretty well functioning and already deployed in the real warehouse/shop system.

SLAM+LIDAR or RPLIDAR is just a bit complex. The provide much more raw data, than needed to get the coordinates and it is very complex task to figure our where you are based on that data. To detect obstacles, avoid them, build the path - it is very good. But to calculate your own location - just a very complex task that, sometimes, returns not so perfect results.

What is FRC? It is unfamiliar acronym to me …

We can track several beacons/robots simultaneously. Current implementation of SW is time division multiple access. Thus, if you have 1 beacon/robot, its data is updated with rate of 16 Hz. Two beacons - 16/2 Hz = 8Hz. Three beacon - 16/3 Hz.
There are ways to make it make it faster, but so far people are playing with 1-2 mobile beacons, so there was no demand.