What Is Sufficient Location Accuracy?

I wanted to post a few things about location accuracy, since that’s one of the key aspects of any location tracking solution.

First of all, you don’t want to buy a gun that doesn’t shoot straight. So you want to make sure the location system produces sufficient accuracy for your use case.

Then again, you can’t have it all: you can’t have inch-level accuracy throughout a multi-floor enterprise campus without investing millions and millions of dollars. Accuracy and cost go hand-in-hand in many cases: If you want inch-level tracking, be prepared to install expensive UWB or such infrastructure, and lots of it. Then again, there are fairly low-cost solutions for achieving very high accuracy in selected, even fairly large areas.

The key is to find a cost-effective solution that solves the use cases you’ve determined.

Accuracy for finding assets and people
For finding assets and people in an enterprise, you can live with a high-quality Wi-Fi network. For example, a Wi-Fi network that’s planned with voice in mind typically has enough access points to support this use case. With some Wi-Fi RTLS engines on the market and a well-designed, dense Wi-Fi network, it is possible to reach a room-level granularity up to 75-95% of the time enterprise wide. This is perfectly OK for a typical asset and people tracking case. If the asset is located on the next room down the hall one or two times out of ten, you will still find them quickly.

However, be careful since the location accuracy between different Wi-Fi location engines varies a lot – even if you have a very dense Wi-Fi network. You don’t want the location system to show the assets on an incorrect floor, or floating in the air far outside the building, for example. Perhaps suprisingly, the “floor hopping” issue is something several Wi-Fi location engines suffer from.

And it’s not just about the number of access points when it comes to accuracy of a Wi-Fi location system, it’s also about the placement of the WLAN access points (APs). If you have 10 APs placed right next to one another, the accuracy will be as bad as with just one AP. Ideally, some of the APs should be placed in the corners of the building, and staggered on different floors. This is not mandatory, but for achieving maximum granularity with Wi-Fi tracking it helps.

If you already have a Wi-Fi network, it simply doesn’t make much sense to me to deploy another infrastructure to find assets or people.

Accuracy for Staff Safety
In healthcare, staff safety is a common application for RTLS. The typical use case is “someone is in trouble ==> panic button/switch is pressed/pulled on the tag ==> an alarm with location information is sent to nearby personnel”.

For this also, tracking with just a good quality Wi-Fi network is typically good enough. Those who assist the person in trouble will find the correct room most of the time, and even if not, it’ll typically be the next room over (and if someone is being attacked, you’ll hear it). But if you want to be sure of the correct room the first time around, you might want to install extra infrastructure in the rooms to ensure high-level room accuracy.

Accuracy for Process Optimization / Workflow Automation
Process optimization and workflow automation to me means optimizing the processes either by analyzing history data, or in real-time, using information about who is/was where, and for how long, etc.

This may require precise information about the rooms the people have been in, and there may be little tolerance for error.

If the error tolerance is low, you want to deploy extra infrastructure to the operating rooms in a hospital, for example, to ensure high accuracy there. The rest of the hospital campus area may be fine with the granularity provided by standard Wi-Fi RTLS.

As said, there are systems out there that do both Wi-Fi RTLS and provide extra infrastructure (such as infrared transmitters) for areas requiring the highest degree of accuracy. Of course, you can mix-and-match different systems together, but the deployment, integration, and system maintenance may become a bit costly and time-consuming.

Huge RTLS Deployment in RFID Journal: http://bit.ly/8U2D6r

There’s a nice article in RFID Journal about the 15,000 tag / 5M square feet deployment.

OSUMC’s director of technology Chad Neal was interviewed also, so it’s not just your average vendor hype. Chad even talks about the evaluation, the accuracy results, the use cases, and so on.

Cisco Live Barcelona is Next Week

Networkers at Cisco Live is happening next week, 25th through 28th of January in Barcelona, Spain. Ekahau reps will be there. If you want to talk about high-performance RTLS that’s optimized for Cisco Wi-Fi, come visit us! Contact marketing@ekahau.com to set up a meeting.

Come see us at Booth #E1.

Cheers,
Jussi

Healthcare Asset Tracking in Florida

Since some of you may not be on the Ekahau news list, I’ve also put summaries of our press releases here. This time, it’s about – suprise, surprise – a healthcare RTLS deployment.

If you have no time to read through all the marketing hype in the press release, here’s a summary: Flagler Hospital, after evaluating various RTLS players (careful evaluation is always a good idea), selected Ekahau RTLS for tracking clinical and IT assets. In addition to various Wi-Fi RTLS players, Flagler also investigated UWB technology as a possible option but decided not go with that in the end.

The hospital is located in St. Augustine, Florida.

As Cisco is the market leader in Wi-Fi by far, and Cisco and Ekahau having a partnership, it’s not a huge surprise that this deployment also runs on a Cisco Wi-Fi network.

> Link to the press release

RTLS Requirements Part 2

Here’s the second part of the RTLS Requirements post. Last week we covered use cases, accuracy, and performance requirements. This time we’ll cover tags, integration, deployment tools, and couple of other points. Lot’s of things to discuss, so let’s get started!

4. Tag Battery Life and Form Factor

The battery life is a key consideration. The more often the tags need to report their location, the shorter the battery life. For tracking assets, Wi-Fi tags waking up once every 10 minutes may be fine, and you’ll get years of battery life. For tracking people in real-time, you may want to update the locations every 5 seconds, but then those tags can be reharged daily or weekly on a tag charger.

The RTLS vendor sales rep can provide you with the exact numbers the system can do. It’s up to you to determine a) what is the minimum battery life I can live with (how often the batteries can be recharged / changed) and b) what is the slowest location update frequency I can live with?  This needs to be considered for each tag type and use case. There’s differences in battery life and rechargeable / not with each RTLS provider, and with each tag model.

There’s tricks that can be used here to improve battery life and system performance: Using a motion sensor to wake up the tag is the most common one. Think of an asset tracking scenario: You can wake up the tag only when it moves, for example, or only when it stops moving. This way, you’ll always know the last location of the asset, and will maximize battery life. Or you can make a combination: “wake up the tag every hour and also whenever it moves”.

To ensure the correct form factor, simply ask for samples of all the tags you’ll use and see if they’re small enough or not.

5. Impact to Facility Daily Operations

When the system is deployed, you want to disturb the operations as little as possible. Cabling is probably the worst: You don’t want to pull more cabling runs unless absoltely necessary. Battery powered readers would be the ones to go with, if possible.

I would highly suggest using software-only solutions and those that utilize your existing Wi-Fi network if possible. But again, no freebies here: if you want true room accuracy (think process optimization), you will very likely need to put some readers in. Some Wi-Fi only solutions, however, are good enough for finding assets and providing staff safety without any additions.

6. Integration with Existing IT Systems

First thing is to ensure that, if you go with a Wi-Fi based solution, that your Wi-Fi vendor is supported. Typically all the major Wi-Fi RTLS vendors are supported by major Wi-Fi RTLS players. If you go with a vendor-specific RTLS solution, bear in mind that you may run into problems if you choose to expand the network with gear from a different Wi-Fi vendor.

The better the system integrates with your existing IT systems and policies, the better of course:

• If the tags need a network connection, ask for at least WPA2 security and DHCP support. If you have an asset management application, perhaps the vendor integrates with that?
• If you want to locate VoIP phones, perhaps the VoIP phone provider has RTLS optimized features (of one or more vendors) built-in to the phone?
• If you have a custom made application you’d like to add location capabilities to, have someone at your IT department (or your application provider or a consultant) look at the APIs from different vendors: How easy are they to integrate with?
• Do you have other Wi-Fi powered devices, like iPhones or laptops that you want to locate? If so, ask for the vendor how that is done, and test that also.

Even though a lot is possible by interfacing with the RTLS APIs, bear in mind: Don’t try to cover all the thinkable use cases in the first phase of the deployment: Keep things simple! Prioritize! Ask yourself, what are the key problems you want to solve today? What will you need in the short-term future? What are nice-to-haves or things that can be left out? Avoid scope creep at all costs!

7. Deployment Methods and Deployment Tool Features

How complicated and pricey is the system deployment? This varies a lot depending on the system: There’s various methods for system calibration, fine-tuning, tag management, the list is long. Ask the vendors for user guides and deployment guidelines – the more cryptic it looks like, the more problems you’ll have and the more time it takes. If you don’t completely understand all the deployment variables, ask. And if the vendor rep doesn’t know them either, get suspicious.

Ask from the reference customers  (you got those from your vendors, right) on how long and complicated the deployment process truly was. A good sign is if the system does not absolutely require vendor presence during deployment (although I would recommend to have them on-site at least for the first days of deployment).

Visual tools with easy-to-understand GUIs are preferred here. Some vendors can provide you with web-based deployment training presentations and even live webinars to go through the tools.

After the deployment, the vendor should be able to provide you with detailed reports about the infrastructure and location accuracy. If it’s a Wi-Fi based system, the vendor should also understand the Wi-Fi network in detail, and thus may be able to provide you with a detailed Wi-Fi site survey report also.

9. Reliability and performance

Of course, every RTLS vendor claims their system is “the most robust and the most scalable and the most whatnot”. Find out about the reliability by asking a few customers who have run the systems for years.

For performance, ask for the biggest deployments the vendors have and call them. An RTLS server should be able to handle at least 10,000-20,000 devices and be able to produce 500-1000 location updates in a second. Your facility size may also affect perfomance: If you have a massive campus, the RTLS server may require more memory. With some RTLS solutions, the accuracy degrades the larger the site is!

10. Requirements for Procedural Changes

When the system is taken into production, people using it and maintaining it will need to change how they work. Examples are employees having to carry tags and leave them into a charger when leaving work and grabbing them again in the morning, or learning to use a new asset management system. This is typically nothing major, but asking the employees about their opinions before the purchase gives you less resistance in the long run – and it’s the nice thing to do.

I hope this helps! Good luck with your RTLS project – may it be a successful one!

Cheers
Jussi Kiviniemi
Sr. Product Manager
Ekahau

The World’s largest RTLS Healthcare Deployment?

Happy new year!

Ekahau just announced a huge RTLS deployment in Ohio: 5 million square feet, and 15,000+ tags to help OSU find stuff and people.

The savings will come from finding assets more quickly, providing better care to patients, improving staff and patient safety, optimizing processes, pretty much everything you can imagine possible with RTLS. Ekahau Positioning Engine location server, and all varieties of Ekahau Wi-Fi tags are used.

I don’t remember hearing of a healthcare deployment that would cover a larger area, or have more tags.

OSU – who carefully tested various Wi-Fi RTLS players before deciding –  is one of the most networked healthcare facilities in the US. They have an extensive Wi-Fi network.  Thus, it was a no-brainer to deploy a Wi-Fi based RTLS instead of installing a separate RTLS infrastructure.  Imagine someone installing readers / sensors and cabling around 5 million square feet…

Location accuracy was another key factor here for the choice: All Wi-Fi RTLS are not the same.

The point: The RTLS market doesn’t seem to suffer from the economic downturn: We’re seeing more and more deployments, which are getting bigger and bigger.

Cheers,
Jussi Kiviniemi
Ekahau