Why tracking in indoor positioning system is unstable

Different indoor positioning systems

There are many different types of indoor positioning systems. And quite a few factors can affect the quality, confidence, and accuracy of tracking. Even more, there are basic mistakes in deploying or using indoor positioning systems. Thus, it is essential to distinguish the problems and apply the “medicine” correctly.

Methods for indoor positioning

What is unstable?

BLE (or Bluetooth Low Energy) is a wireless personal area network technology designed, first of all, for data transmission.

BLE is not designed for positioning. Like in any RSSI-based (radio signal strength indicator) system, for example, WiFi, positioning is just a nice by-product for them. Yes, BLE does data transmission between phones, phone, and hands-free, etc.; it can also be used for tracking. In addition. With serious trade-offs.

The fact that RSSI measurements and positioning is a just a by-product for RSSI-based systems (BLE, WiFi) is their strongest point, and it is their weakest point.

Strongest:

You don’t have an additional mobile beacon to track your location. Typically, your mobile phone or tablet or similar gadget already has a BLE module and by adding special SW/app to your device you can track your mobile phone/tablet/gadget

Weakest

RSSI is a very poor indicator of the distance. Only because the software of the indoor positioning system is so smart, the system may somehow determine the location from the very jumpy RSSI readings. But even with the best tricks you shall realistically expect to know the location with 2-5m accuracy, if beacons are placed every 10-30m

BLE can be more complex and more advanced and support Angle of Arrival (AoA) and Angle of Departure (AoD)

  • Combination of imprecise RSSI + angles => up to 3 times better accuracy of positioning
  • Multiple antennas =>  larger size, larger cost, limitations of applicability
But the confidence of tracking improves as well as the accuracy. It can reach even 1m in maps with relatively densely populated beacons.

When to use BLE for indoor positioning system

When you need a room-level accuracy and when you already have a mobile device with BLE onboard, the BLE RTLS is a good choice. This is an ideal case. Typical applications:

  • To find a gate in the airport
  • To find a masterpiece in museum
  • To find a shop in a shopping mall

 

You place BLE beacons every 10-30m inside your premises. The beacons are small and can easily camouflaged. They emit Bluetooth signal once per second, for example, and with this location update rate can run on internal battery for 1 year or so. After 1 year you just change the battery on all of them.

 

The mobile phones that you use as your mobile trackers don’t need any changes or add-ons. They run an application that communicates with stationary beacons and calculates its own position based on the radio signal strength to particular stationary BLE beacons and known position of those beacons. Everything works very well albeit not too precise – 2-5m typically, which is OK for many applications (see examples above), but not OK for the majority of industrial applications.

Why exactly BLE is not good for industrial applications

Insufficient accuracy

Typical requirement of people tracking for industrial applications is better than 1m. Often – better than 0.5m. BLE is simply not capable in practice to provide such an accuracy or you need to place the BLE stationary beacons very densely – every 2-5m.

Metal around

When we receive requirements about an indoor positioning system, very often potential customers specifically highlight that they have metal around. We immediately know that they have burnt their feet with RSSI-based systems, most likely – BLE. Why? How do we know?

Well, as already discussed above, the RSSI-based indoor positioning systems use radio signal strength from a stationary beacon to the mobile beacon to estimate the distance. Yes, the farther your mobile beacon (phone) is from your stationary BLE beacon, the weaken the signal shall be. In practice, however, the RSSI fluctuates so drastically that the same RSSI may be measured on 2m and on 10m. What distance is then?!

When you have nice empty space of airports or shopping malls full of radio transparent glass or thin walls, everything is more or less OK. But when you want to go to a very packed warehouse full of metals shelves, palettes, forklifts or other heavy machinery, the RSSI picture changes so drastically, that the BLE-based tracking simply does not work with acceptable accuracy. Like really doesn’t work. You can be on this aisle or another one or another one. And what is the point in the system if you can not find what you want to find?

RSSI picture changes all the time

There is not only metal in the industrial environment, but moving metal. Imaging, you are not moving, but a forklift is moving 5m away from you. But your location measured by the BLE-based system jumps. Why? Well, it happens because of multipath fading. Thus, a system calibration that you have done 1h ago may become very invalid right away.

Of course, there are very smart algorithms that try to somehow recover, re-calibrate the system. But at the end of the day, it does not work well in the real industrial environment. RSSI-based systems are simply not designed for that kind of environment.

You have to have dedicated mobile beacons in industrial environment

When you track mobile phones, life looks cool: download an app and you are tracked. But what you need to track not a phone, but something else? Helmet? Crane? Palette? Forklift? – you need to install a mobile beacon on it. Just another device. Well, for precise indoor positioning systems you always need a mobile beacon, but they precise at least. With this situation you have both (1) poor accuracy and (2) additional mobile beacon with all difficulties of charging, additional costs, etc. One of the strongest advantages – not requiring a separate mobile beacon for tracking – disappears from BLE systems in industrial environment.

What is better than BLE for industrial applications?

There are two major technologies recommended for industrial indoor positioning systems:

Read more: What is better than BLE for indoor positioning system?

As discussed in Review and comparison of indoor positioning technologies and methods with focus on industrial applications, both technologies rely on time of flight – not on RSSI. That is the key advantage, because it does not matter what is around: metal, concrete, wood, or nothing at all. UWB – time of flight of radio wave. Ultrasound – time of flight of ultrasound. Since the speed of sound is a million time slower than that of radio wave, ultrasound based solutions are inherently far more accurate than UWB.

Therefor, the key practical advantages:

  • If you need the highest accuracy available (at least 10 times better than UWB), choose, Marvelmind Indoor “GPS”
  • If you have wooden/glass radio transparent walls or your environment very acoustic noisy then choose UWB
  • If you need a lower-cost solution – choose Marvelmind Indoor “GPS”

 

Of course, like for any precise indoor positioning system, line of sight is a must, but the definitions of what exactly “line of sight” are different:

  • Radio transparent walls (thin glass, wood, etc.) give line of sight for UWB, but not line of sight for ultrasound
  • At the same time, typical “breathing” cloth is transparent for ultrasound as well, even if the mobile beacon is not visible, i.e. out of line of sight – for light, not for ultrasound

What about optical tracking or LIDARs?

Well, optical indoor positioning is not good in practice:

  • Optical systems can somehow work for monitoring, but not really for positioning and even worse for navigation
  • Ultrasound stationary beacons can run on battery, though fixed power supply is recommended. The power supply in this case is just a basic electricity grid (~110/220V) or a USB or +12-36V – whatever is easier available. UWB typically requires fixed power supply for stationary beacons (anchors); often the anchors also need Ethernet for data exchange and synchronization. But cameras for sure require both power and high-speed connectivity backbone – either Ethernet or wireless. Also, servers to process massive data stream. It is simply costly and cumbersome to provide all that connectivity
  • Line of sight is really a problem
  • Accuracy depends on the distance unlike in the time-of-flight systems. Higher resolution cameras become even more expensive and have other drawbacks
  • Typical industrial environment is poorly lighted – either too dark or too bright or both – too high dynamic range
  • In many cases, dust, fog, temperature change – condensation

LIDARs are not suitable for the majority of industrial positioning systems:

  • In fact, it easier to mention where the LIDARs can be OK for industrial indoor positioning systems – expensive AGVs and expensive robots. In general, LIDARs are good for obstacle detection – not for positioning
  • LIDARs are not suitable for people tracking at all

Scroll to Top