Blog skyhook logo It's Time to Solve LPWAN's Location Challenge with TDOA and Wi-Fi

Oct 2, 2019 1:14:08 PM

It's Time to Solve LPWAN's Location Challenge with TDOA and Wi-Fi

Posted by Kipp Jones

Read All Posts

This post originally appeared on IoT Agenda. This is the first in a three part series on IoT, Location, and TDOA. Read Part two and Part three

The internet of things continues to gain momentum and in so doing is creating new opportunities and innovation in both technology and business use cases. Many of these new use cases are demanding smaller and smaller devices that have less room and money to support the need for a large battery. At the same time, many of these same use cases are of the ilk that requires less data transmission to support the mission. Use cases, like being able to track a single pallet of goods around the globe reliably or the ability to track rental bicycles around town, are demanding technologies that heretofore did not exist.

The following describes some of the consequences and responses to these customer demands, both in terms of new network capabilities and in terms of location technology to support these demanding requirements.

The creation of low-power wide area networks

The internet of things is driving connected devices to be smaller and cheaper in order to enable many different systems for businesses, from indoor asset tracking to global supply chain logistics. The problem is that these smaller things still need batteries, and batteries can end up defining the form factor and other attributes that can be realized (e.g., weight, size, cost, life of device and so forth.). The drive to shrink the battery while extending the life of these devices requires tradeoffs and continued innovation. One activity that uses this limited power is network communications.

This need combined with the more limited bandwidth requirements for many IoT deployments begot the invention of low-power wide area networks (LPWAN). These new networks have special-built protocols that have fundamentally different priorities than today’s data-voracious mobile consumer devices. By tuning a network to provide long-distance, low-power, low-bandwidth communication, you have a set of capabilities that both matches the primary operating needs of a large set of IoT use cases, as well as one that reduces both the cost and the power consumption of the devices — extending the life, shrinking the size and lowering the cost of the devices and their operational overhead.

These new LPWANs come in several different flavors, from more open, purpose-built networks such as LoRaWAN to more proprietary purpose-built networks from Sigfox to 3GPP standards-based cellular networks designed to solve this same set of problems, including LTE Cat-M and LTE NB-IoT. Each of these networks has varying properties with respect to bandwidth, costs, network coverage, ability to provision private networks, roaming and, of course, power consumption.

While all of this sounds great, many of these networks are still in deployment phase around the world. Yes they are available, but at different scale and coverage depending on the network. The decision tree for determining the best network choice is nontrivial and constantly changing as the competitive offerings are continuously in flux.

IoT and location

But, let us not forget that nearly all IoT use cases require some element of location. The required accuracy and availability of device location varies from case to case, but very rarely do we see deployments in which location doesn’t play an important part of a complete system.

As with other devices, including your smartphone, there are only a few use cases where a single technology will give you the level of location accuracy and availability that you need. While many people assume that GNSS solves that problem by itself, it really doesn’t — it doesn’t in your smartphone and it won’t in most IoT cases either. GNSS lacks certain capabilities, like the ability to position you indoors. Yes, it sometimes works, but oftentimes it doesn’t — so if you need reliable location indoors, you better be considering other alternatives. Oh, and remember that power thing? GNSS can be one of the more power-costly options when it comes to providing location for a device. Combine this with the need to reduce the overall bill-of-material costs for these devices and you will find many IoT companies opting to build devices without GNSS.

So what is a person to do? One nice thing about LPWANs is that they are able to provide a location on the network side based on the timing or strength of signals reaching a set of their communication gateways/base stations.

Since LPWANs provide wide area coverage at low data rates, the base stations can be sparsely deployed. This naturally has negative impact on network-based location accuracy, especially when it relies on signal strength (power) measurements to estimate range from device to base station.

LoRaWAN networks natively support timing measurements that allow it to use a technique called time difference of arrival (TDOA) to estimate distance and compute the location.

TDOA was developed and patented by Skyhook for locating mobile phones and devices more than 20 years ago. Since that time, we have deployed two nationwide TDOA systems in the U.S. and in multiple international markets using a variety of air interfaces and frequency bands. This experience has provided valuable knowledge and feedback from real-world large-scale deployments. As a result, the company has continuously enhanced and refined our software and tools for mitigating multipath and otherwise improving system performance in varied environments.

In the years since, TDOA has been adapted to support native LoRa TDOA locations for IoT. It can process gateway time stamps and act as “solver” to produce optimal fixes based on triangulation of isochrones.

TDOA location can be accurate enough for some IoT use cases, at which point you may be done. However, accuracy can also be heavily reliant on the density of LPWAN base stations deployed which, as mentioned, varies greatly today by geography and morphology (urban versus suburban, etc.) and thus may not be sufficient to support your use case in all areas.

Why Wi-Fi?

One popular method used in smartphones is to pair Wi-Fi with GNSS and/or cell for positioning and assistance data. Using Wi-Fi as a second source for location does a number of things to enhance the system:

  1. It provides indoor location capability and accuracy;
  2. It provides improved location performance for GNSS;
  3. It provides more accuracy than cell-based locations (including TDOA on LPWANs); and
  4. It can be combined with other signals (hybrid) to create an improved location system than over any single choice on its own.

In fact, the last option is what we recommend for LPWAN deployments. First, by adding Wi-Fi location to the system, you extend the reach of your devices and provide a higher precision location technology. Second, by acquiring a Wi-Fi scan and sending that to the network for localization, the burden on the battery is minimized. Third, by combining Wi-Fi with LPWAN network-based location, we can improve the overall location accuracy and reliability across the network. Lastly, Wi-Fi can be used instead of GPS for low-cost devices without a GNSS chipset where native LPWAN location is not sufficiently accurate.

New call-to-action