A Non-Technical Manager's Guide to Indoor Tracking Systems
Or, Dazzling Your Technical Staff with your In-depth Knowledge

By Adrian Jennings, CTO Time Domain Corporation

Introduction

If you read past the title, then you are probably already considering installing some kind of tracking system in your facility. You may wish to track assets, staff, patients or any combination of the three. If that’s the case, then you’re probably also confused – there are simply so many options out there, and so much marketing hype, that selecting the right technology for a pilot or full installation can be daunting. System Integrators can help sort through the confusion, but at some point you are the one accountable, and you want to feel comfortable with the technology you are about to spend part of your capital budget installing.

Understanding tracking systems isn’t so hard, in fact you already know everything you need to, and you just need someone to explain how it relates to tracking systems. That’s what this paper attempts to do: connect things you already know about physics (yes, you know lots about physics) with the way tracking systems work.

If I know so much about physics already, then how come I’m so confused?

A very good question indeed. The trick is having someone point out what you didn’t know you already knew.

First, we need to start with some basics. We’re going to start to develop categories of tracking systems so that you can apply what you know about one system to others of the same type – you’ll be surprised at how few types there actually are.

So, how do tracking systems work? First of all, they all rely on some kind of installed infrastructure, and some kind of tag. The installed infrastructure consists of devices fixed to walls and ceilings throughout the building, in much the same way as wireless networks. Some tracking systems need a lot of devices, some few – we’ll come to that. We’re going to call these devices “readers” for the purpose of this paper.

The other critical piece of any tracking system is the tag. This is usually a small device which is mounted on equipment or worn by personnel to be located. The tags send out a signal which is heard by the readers. The readers each determines how far the tag is from them, either by how loud the signal is, or how long it takes to arrive. The results from multiple readers in known locations are then used to calculate the final 2- or 3-dimensional location of the tag.

That doesn’t seem so complicated….

It’s really not.

The complication starts with the large number of ways that people have found to implement such a system, each of which having its own pros and cons. To understand the differences, you need to understand what it’s like to be a reader: what is it like to listen to tag transmissions, and what kind of transmission makes it easier to figure out how far away the tag is. These questions are crucial to understanding how a tracking system will work in real life.

This is where you already now a lot more about physics than you think. Consider the last time you were in a thunder storm: In a storm, we’re taught to count from when the lightning flashes to when we hear the thunder in order to estimate how far away the storm is.

To some extent you can figure out the distance by listening to how loud the thunder is, but that only really gives a coarse indication: my neighbor’s house is 500 yards away, but something about the landscape funnels his country music till you would swear that his outdoor speakers are on my deck, not his. In order to be more accurate we assume that the flash of lightning reaches us instantaneously (a pretty good assumption given the speed of light), and we use that as a timing reference to figure out how long it takes the sound to reach us.

That’s pretty much what it’s like to be a reader. It gets a timing synchronization signal (either over a wire or over the air) and counts from there till it hears a tag, just like counting seconds between the lightning and the thunder. By measuring the difference in the time of arrival of a pulse at three different readers, the location of a tag can be pinpointed in two dimensions. (We know which direction thunder is coming from because we have two ears (readers), but not the exact location because that would require a third ear.)

Here’s some more physics that you know: it’s a lot easier to pinpoint the time a clap of thunder arrives if it’s, well, a clap of thunder. By “clap” we mean a very short burst of sound like the cracking of a whip – one of those mighty thunder claps that rattle the house. It’s much harder if it’s one of those long rolls of thunder, building slowly to a crescendo then dying away. When exactly did the sound first arrive? Did you miss the quiet start because you weren’t paying attention? Was the start of the thunder even too quiet for you to hear, giving you no chance to time the exact start of the sound? These same problems plague readers too: short pulses help to determine accurate arrival times, long pulses are harder to judge. The result is that long pulses make the timing, and therefore the location determination, inaccurate; where as short pulses are very precise.

Pulses of what, exactly?

Here’s where we start to get into the different kinds of tracking systems available. There are two types of pulses that can travel through the air from a tag to a reader: sound and electromagnetic pulses. There are many kinds of electromagnetic pulses (X-rays, gamma rays etc.) but the ones used in tracking systems are limited to radio frequency (RF) and infrared light.

As we’ve said, all that really matters about the pulse is that it is very short – we need a clap of thunder not a peal. OK, that’s not actually all that matters – we’ll come to some of the other considerations soon. None of the tracking system vendors are telling you their pulse length though, are they? They want to tell you about the frequency they operate at – so how is that helpful?

Well, as you know, the width of a pulse in the time domain is inversely proportional to its bandwidth in the frequency domain. You didn’t know that? Sure you did – here’s another piece of physics that you didn’t know that you knew:

When you’re tuning in to a radio station, you turn the dial to select different frequencies until you find the right one. Only not the right “one” because often you can move the dial left and right quite a bit and stay on the same station. That’s because the radio station has “bandwidth” meaning that its signal spreads across a number of frequencies with a certain width.

Now, let’s go back to the thunder storm, only let’s focus on the lightning now. Lightning always comes as a flash of light of very short duration. We say that it is “narrow” (short) in the “time domain” (meaning when we measure its duration). But what does lightning look like in the “frequency domain”? In other words: how does it show up on a radio dial?

When you’re listening to the radio in a storm, you can “hear” the lightning as a burst of static. Now ask yourself this: which station can you here it on? NPR? Classic Rock? Gospel? The answer is all of them – the burst of static can be heard on all frequencies – it has very wide bandwidth in the frequency domain. So, a pulse which is short in the time domain has a wide bandwidth in the frequency domain. Go back and read the italicized statement three paragraphs up…. As I mentioned – you know a lot more physics than you thought….

OK, this is all very interesting, but weren’t we talking about tracking systems?

Indeed we were. Let’s recap: Thunder teaches us that you can measure only how loud the pulse is, but that’s not such an accurate indicator of range. Thunder also teaches us that it’s much better to have a short pulse than a long one if you want to measure the time of arrival of a pulse.

Lightning tells us that short pulses have wide bandwidth when we’re thinking about frequencies, and that gets us one step closer to deciphering vendor doublespeak when choosing a tracking system.

How close? Almost there. We now know the bandwidth of a tracking system has a big effect on its accuracy, but there’s a little more to bandwidth that really sets apart systems that work well in the real world from systems that only work well in the laboratory.

Let’s think about thunder again. It’s fairly easy to count the timing of a single clap of thunder, but what about two claps? What about two long peals? What if they overlap, so that you can’t hear the start of the second peal because the first one is still rumbling? Now you have a real problem – you just can’t figure out the timing of the second peal (pulse). Now we’re back to short pulses: if the thunder comes in short claps, then there’s a much reduced chance of them happening at the same time and swamping each other.

Overlapping pulses are a problem when you have lots of tags, so short pulses help you to track many tags without them all interfering with each other. But it gets worse that that: we’re trying to track tags inside a building and the problem with that is, well… the building. It has walls – lots of them, and each one reflects the pulses from a tag. A reader doesn’t hear a single pulse from a tag; it hears the same pulse many times over, echoing off walls, furniture and other objects in the building. Now you REALLY need short pulses – if they’re long to start with, all the bouncing signals will overlap and obscure the one you’re trying to hear – even if only one tag is present!

Bouncing pulses? I thought they went through the walls….

The fact is that some of the pulse bounces off the wall and the rest if it goes through. How much goes through? The answer to all good questions is: it depends. Specifically, it depends on the physics of the wall, but you know all about that already, right?

Well, actually you do. Anyone who’s been kept awake late at night in a hotel room by the TV next door knows that the sound that comes through is booming, and muffled, not squeaky and high pitched. Why? Because walls let low frequencies through, and bounce high frequencies back.

This is important: if the frequency is low enough, there won’t be any bouncing of pulses to confuse the reader. Additionally, the readers can be spread apart because the pulse will happily traverse many walls to get to them – think how many readers you’d need if the signal didn’t go through any walls at all….

OK, I get it: I need high bandwidth and low frequency. Where do I sign?

There’s bad news of course – there always is. It turns out that life is all about compromises. There are two things that limit the bandwidth of a system: the law, and practicality, and you can’t get around either (until recently).

The law is that of the FCC, which specifies the bandwidth that a system can occupy at any given frequency. Historically, the radio frequency part of the spectrum has been more about things like radios than tracking systems, and only narrow slivers of bandwidth are available at any given frequency (until recently). The practicality issue is that it becomes very difficult indeed to make systems with very wide band- width at very low frequencies. The result is that, in general, high frequency systems have high bandwidth, and low frequency systems have low bandwidth (until recently).

OK, I’ll bite: “until recently”?

There’s a new technology in the tracking business that breaks the practical limit of frequency and bandwidth. It’s called Ultra Wideband (UWB), and its bandwidth is, in fact, ultra wide. New FCC rules allow UWB to operate over unprecedented bandwidths, thousands of times greater than conventional technologies, enabling a new generation of precise tracking systems because of the correspondingly tiny pulses.

The key advantage of UWB is that it can generate enough bandwidth for very precise tracking, at frequencies low enough to penetrate walls. The ultra wide bandwidth means that it uses ultra short pulses, and these in turn give ultra precise location, in ultra-reflective environments, with ultra high tag densities.

And you can appreciate the value of UWB because now you understand the physics.

There’s more to say about UWB – the advantages don’t end there, but first let’s wrap up what we know, and then see how the technologies compare.

That’s right: there was mention of tracking system categories earlier in the paper.

Indeed, and we’re ready to understand them now. Here are the main categories:

1. There are acoustic systems, and electromagnetic systems.
2. Electromagnetic systems come in infrared (IR) and RF varieties.
3. RF systems come in UWB and conventional varieties.
4. Conventional RF systems come in narrowband (low bandwidth) and spread spectrum (medium bandwidth) varieties.
5. Spread spectrum systems come in signal strength and time of arrival varieties.

With the added knowledge that acoustic (ultrasonic) systems and IR systems both operate at extremely high frequencies, you are ready to create the comparison table (see above):

The advantage of a UWB system is clear from the color coding, in that it combines the accuracy and tag capacity of very high frequency systems with wall penetration closer to that of medium frequency systems (low frequency systems just aren’t used for tracking – not enough bandwidth down there).

Why do you care about wall penetration again? Because you don’t want to have to pay for three readers in every room to get accurate 2D location when there are systems that can hear tags through walls, vastly reducing the number of readers required (and therefore the cost of the system). Some of those tracking system categories still don’t sound familiar to me….

Let’s wrap this whole thing up with some context around what people are advertising, and how that relates to the different categories. For each category we’ll briefly discuss the industry names for the different systems, and some additional characteristics of each. The red, yellow and green boxes from the table above are recreated to give an indication of pros and cons; a short “What it doesn’t say on the box” section for each technology points out the things you really need to know. We’ll conclude with a final “Don’t believe the hype” discussion that highlights some key questions you should be asking your hardware vendors.

ULTRASONIC

Ultrasonic systems might also be called “acoustic” or “sonar” – it all means the same thing. The pulses are sound waves just above the audible range of frequencies. These systems use time of arrival techniques with extremely short pulses that can provide excellent precision. The down side (and hence the red blocks) is that there is no wall penetration at all, so that a 2D tracking system would require three readers in every room. Since that would be cost prohibitive, these systems are generally offered in a one-sensor-per-room configuration, telling you which room an object is in. Room level accuracy only? That rather defeats the advantage having all that bandwidth….

What it doesn’t say on the box: Ultrasonic systems suffer from a good deal of noise in the environment (there’s a lot of ultrasound around – we just can’t hear it). Ask your vendor how well it does in noisy environments. Try rattling a handful of coins near a sensor and see if the system can cope.

INFRARED

Infrared systems will normally be called by the abbreviation “IR”. This uses technology just like your TV remote control to measure the time of arrival of pulses of light, just outside the range that we can see. (Your low-light video camera can see it – ever pointed your TV remote at your camera in night mode and recorded what happens when you push a button?) The comments under Ultrasonic apply here too, and IR systems are generally also used in a one-sensorper- room configuration.

What it doesn’t say on the box: IR signals are blocked by just about anything, leaving dead spots in a room. Ask your vendor how they fill in the dead spots, and make sure you test the system thoroughly.

ULTRA WIDEBAND

Ultra Wideband systems will often be known by the abbreviation “UWB”. These systems use extremely short pulses of RF energy at very low transmit power to measure the times of arrival of pulses. The very low transmit power has the added advantage of being completely safe around people, and practically invisible to any other wireless systems in the vicinity (e.g. your WiFi communications network). With the best combination of precision, wall penetration, tag capacity and echo immunity, UWB scores the most greens and yellows of any system (and no reds).

What it doesn’t say on the box: The very low transmit power limits the range of the readers to operation through two or three walls. This is better than ultrasonic and IR systems, but drives a higher reader density than spread spectrum and narrowband systems. Ask your vendor about proximity-only modes to decrease the number of readers where you don’t need precise 2D tracking (you can add more later if you need to).

NARROWBAND

Narrowband systems mostly come under the description of UHF (the same UHF that you’re used to from passive tags, but this time we’re talking only about active tags). These systems score the most red chiefly because of the very low bandwidth which affects accuracy, tag capacity and echo immunity. They do have the best wall penetration though, being the lowest frequency systems offered for tracking applications.

What it doesn’t say on the box: UHF systems are well known and therefore without too many pitfalls. Do be wary of claims of great accuracy though – where did they get all that bandwidth in those crowded spectrum regions?

SPREAD SPECTRUM – SIGNAL STRENGTH

You won’t hear anyone say “spread spectrum”, instead they’ll say “2.4 GHz” (“gigahertz”) or sometimes “2.45 GHz” but most often “WiFi”. WiFi systems are the only ones to consider here, since they reuse your existing wireless network infrastructure, cutting down on your initial installation cost. The accuracy is not very good, and the systems suffer from very poor or no performance in wide open spaces or outdoors. The chief advantage is the cost saving of wireless infrastructure reuse.

What it doesn’t say on the box: Sure enough you can reuse your existing WiFi infrastructure, but you’ll need to add more access points: data networks don’t have the necessary density of readers (called access points in the WiFi world) to support location services. If you’re already running voice-over-IP then you are in much better shape. Ask your vendor how many readers you really need.

SPREAD SPECTRUM – TIME OF ARRIVAL

Again, these systems will be called 2.4GHz, 2.45GHz or WiFi. They score well, with no reds, but less greens than you would like. WiFi systems use the same kind of signals as your wireless data network, but the readers are not the same as your existing access points – you’ll have to buy all new for the tracking system. Access points which handle your data network as well as time of arrival tracking are coming onto the market, but you won’t quite get the same reuse as with a signal strength system.

What it doesn’t say on the box: Even though TDOA WiFi is the “cream of the conventional crop” it still has its problems, most notably in indoor environments. Ask your vendor about performance in areas with many walled rooms, and don’t be surprised if signal strength WiFi is all that can be used here.

And Finally….

Don’t believe the hype.

Yes, it’s true: there’s a lot of hype in the marketplace, and sometimes it’s hard to get to the truth before you run a costly pilot program. Here are a few pointers to help you make good choices.

1. Make sure you know what you need your tracking system to do for you. Don’t just think about your current pain points; think ahead a little for additional usages. You’re going to like having a tracking system, and you don’t what to face a costly upgrade (or worse, a full replacement) when you start to use it for additional applications. You’ll save money in the long run if you think beyond your current needs.
2. Choose a system with flexibility. You won’t be able to predict all the uses that you’ll put your system to, so you need a technology that can grow with you. Ask your vendor whether the system can operate in different modes. Is there a low cost option for identifying which areas an item or person is in? Can you apply high precision only where you really need it? Will the same tag work in both zones? Two different systems won’t help you much if you have to tag everything twice….
3. Try to make an apples-to-apples performance comparison. You don’t want to know the range of a reader in the laboratory, in the parking lot, or on the dark side of the moon: you want to know its range in your building. Make sure the different vendors give you performance predictions in that same context, otherwise their numbers are meaningless. Do I need to mention not to believe the marketing hype? Ask you vendor if they’ll bring a reader to your building to show you the performance they’re promising.
4. My favorite performance parameter is tag battery life, where the specifications differ wildly. Here’s one where you definitely need to ask some supplementary questions in order to compare technologies. The easiest way to conserve battery power is simply not to switch the tag on, and many vendors will quote battery life with a tag operating at once per day or longer. That might be fine, but if you want to track people moving around you’ll need the tag to be operating at more like once per second. Make sure you know what kind of things you want to locate, and get battery life figures accordingly. Also, you need to know that most batteries only have a shelf life of 4 to 5 years (meaning that they’ll run themselves flat in that time even if they sit on a shelf unused). Anyone quoting a tag battery life longer than 5 years needs to explain what special battery or technique they are using to achieve this.
5. One technology is going to work more or less the same as another technology in the same category. Are your UWB vendors giving you wildly different predictions? Your WiFi vendors? Nobody underestimates – try to pin the overachieving vendor down: why is their system so much better? What breakthrough have they made? Breakthroughs happen, and sometimes a technology leaps ahead, but make sure the spec sheet isn’t the only thing that’s leapt ahead.

Are we there yet?

Yes, that’s it. You now have an overview of the different systems out there, and how they stack up against each other. You are armed with some general tracking performance questions, and some specific technology questions to really make your vendors give realistic answers. You also have some intuition about the underlying physics, which arms you with confidence when talking to technical people, and hopefully was fun to learn.

That’s it. Happy shopping and call if you have any questions: we’re here to help.

DOWNLOAD WHITEPAPER

To download a printable PDF version of this whitepaper, click here.