Shop
0

How to increase location update rate?

Quick and dirty

  1. Set the Location update rate to 16+Hz in the Dashboard. The system will reach the update rate ceiling based on other limitations
  2. Disable Realtime Player to reduce an additional latency due to it to zero. Set the Window of averaging = 0 and Distance filter = 0. It won’t increase the update rate but will reduce the latency
  3. Reduce the number of mobile beacons in NIA, if feasible. The update rate will improve proportionally
  4. Reduce the number of mobile beacons in MF NIA to less than 8, less than 16, etc. The update rate will improve proportionally to int(N/8)
  5. Define service zones. If you don’t, the system uses the default service zones of 30m, and the update rate may not be optimal. If you reduce from 30m to 10m, for example, your update rate may rise up to 3x, if it was limited by the size of the submap, which is often the case
  6. Change from default 153kbps radio profile to 500kbps. If it was radio limited, you might see an increase. Usually, it is less than that with service zones or zero if the system isn’t radio-limited. But becomes very important in swarm robotics and IA. More about radio connectivity
  7. Choose the location update rate on the modem vs. the location update on the mobile beacon. If you update the location of every mobile beacon on the modem every time, the location update per mobile beacon will be lower due to higher radio traffic

Detailed explanations

There are several key parameters in any real-time locating system (RTLS) or indoor positioning system (IPS), or indoor navigation system (INS), which are just different names for the same thing. Among the key parameters:

  • Cost (HW + SW + deployment + running, etc.)
  • Accuracy/precision
  • Location update rate and latency
  • The number of mobile objects tracked

There are many more, and you can learn about them in detail: https://www.youtube.com/watch?v=zg3oW_U_jdY&t=3949s. Let’s discuss deeper the two of them more:

Accurate RTLS for swarm robotics

Notice that though location update rate and latency are usually linked, they are not the same thing. For example, you may have a high update rate and high latency simultaneously.

Also, remember that the accuracy of position measurements doesn’t depend on the location update rate. At the same time, if your system can tolerate more latency in the position measurements, it is possible to achieve higher accuracy of measurements. See more: Does accuracy depend on the location update rate?

Location update rate

Location update rate – how many times per second the location has been measured.
For some customers, “fast” means once per minute, for example, for static assets tracking – palettes, etc. For some other tasks, even a 100Hz update rate is barely fast enough – VR/AR and similar.
We support up to 30-40 Hz for very small submaps of 2-3m. Typical update rate with default settings with average-sized submap of 10-15m – 8Hz. For larger maps of 30m and default settings – 6Hz or so.
If you don’t specify the service zone, the submap assumes the largest submap, and you get 6Hz or so.
Notice that we have both location and direction for the Paired Beacons configurations. While the location is measured with a typical update rate of the ultrasound system – 8Hz or so – the direction is calculated based on sensor fusion with IMU. The sensor fusion gives 100Hz update rate for direction and ~10ms latency for it.

Reduce latency

Latency – a delay between the moment the location has been measured and the moment we know the calculated coordinates.
 
Latency is usually a more complex task than update rate. For example, the Realtime Player is a relatively simple method, but it increases the update rate x10..x100 times. To do the same with latency, an architecture or even the underlying positioning technology could have to be changed.
 

There is a clear trade-off between latency and accuracy of tracking. By applying more filtering or averaging, it is possible to achieve better tracking quality, but latency will suffer. Thus, each case must choose between accuracy/quality and latency.

More on the adjacent subjects:
– Does accuracy depend on the location update rate?

IMU+Ultrasound Sensor Fusion Post-Processing

To have real 100Hz tracking, use IMU + Ultrasound Sensor fusion.
 
 

Further development

If the current 8-16Hz location update rate and 100Hz for direction for typical configurations are not enough, what are the options for further development?
  • Please send us a request to info@marvelmind.com. Maybe, you don’t need the high update rate or low latency as you think you are. Perhaps, there is another solution
  • For swarm drones, MF NIA is the best option now. But for more than ten drones, the update rate of <4Hz may already be not good enough. You may need to change to IA instead of MF NIA (see more: https://marvelmind.com/pics/architectures_comparison.pdf). Since mobile beacons in IA receive an ultrasound, you could have to do the following:
a) reduce the size of the submap and increase signal/ration
b) install special shields/screens protecting microphones from the noise of the drone’s noise can help
c) limit the relative locations of the drones against each other. For example, they can fly side by side but not above each other – at least, not too close, to keep a sufficient signal/noise ratio
  • Sooner or later, we will complete Ultrasound+IMU sensor fusion. With an ultrasound location update rate of more than 2-4Hz, the sensor fusion with IMU will kick in and provide a 100Hz update rate and 10ms latency for location. It is similar to what we already have for direction and location. The key step in this development is 10ms because, with the help of the Realtime Player, it is already possible to have a 100Hz location update rate, but with 1-2sec latency – not with 10ms latency. So small latency requires sensor fusion with IMU, which is very complex for the general case. Please send us an email to info@marvelmind.com because the task may be less complex for your case with unique constraints.

If anything is unclear, contact us via info@marvelmind.com