The proliferation of handheld positioning needs no introduction.
Consumer reliance on mapping and location services from handheld devices is now so entrenched that urban dwellers who have navigated a city their entire life will fire up their Google Maps or Waze app, mount their phone to the dashboard and do as they’re told. Consulting such services is a daily occurrence for most smartphone users.
Satellite-based positioning is a mature technology. GPS, or any of the GNSS-based networks, is far and away the most popular positioning technology – highly reliable; accurate; relatively low-cost, and easily deployed for users. However, GNSS falls down when it can’t receive frequent, high-quality satellite signals, a familiar scenario inside most buildings and in ‘urban canyon’ environments.
New solutions are needed indoors. A vast constellation of different indoor positioning techniques are out in the field, evolving, maturing and desperately fighting to gain a foothold in the market.
These techniques and technologies have been silent achievers, rapidly developing and fiercely competing in their own little niche, slowly stealthing into the mainstream. Tech giants Google and Apple have been quietly integrating technologies pioneered by smaller players into their products as they continue to build out their location services and provide new ways for locationally-challenged smartphone users to find their way around the built environment.
WiFi is by far the most common form of network leveraged currently leveraged for indoor positioning services, followed by Bluetooth Low Energy (BLE), and Radio Frequency (RFID), usually utilising approximate, Cell-ID, range-based or fingerprinting positioning techniques. (See sidebar for a detailed breakdown of indoor positioning tech heavy hitters)
With little fanfare, Google began mainstreaming these technologies as they launched indoor maps in 2011 for the U.S. and 2012 for Australia, utilising WiFi network hardware. Apple acquired WiFiSlam for $20 million in 2013, an ingenious start-up with a method that collects anonymised data from WiFi adaptors and other sensors in a users’ devices, and using this to create a WiFi fingerprints database.
Apple’s iBeacon and Gimbal’s proximity beacon are typical examples of commercial Bluetooth beacons, using the approximate positioning technique to acquire the mobile station’s location. The positioning accuracy of these technologies varies from several metres to, up to tens of metres.
Light the way
The maturation and spread of LED lighting has provided an unlikely and unique technological opportunity for indoor positioning systems. Signals transmitted by LEDs can
be used to determine the position of a mobile station, based on Cell ID positioning. With specifically designed LED transmitters and receivers, time-difference-of-arrival measurements can be obtained, and accuracy of up to several centimetres has been recorded.
Philips is leading the charge in championing this technology with the retail market in mind, embedding signals in the light wave modulations emitted by LEDs, building Bluetooth services around their system, and tapping Microsoft to develop these services further. Other players in the space, lighting giant Osram and startup Goee, are approaching more cautiously with a Bluetooth-only system.
Satellites on the ground
Locata Corporation is a Canberra-based firm that landed a fairly high-profile customer in 2015 – NASA. Locata was chosen to install a ground based positioning network at their Langley facility in 2015. Locata has built a patented system around pseudolites – a contraction of ‘pseudo-satellite’ – and perhaps the closest thing possible to a terrestrial GNSS network. are ground-based transmitters that employ a signal structure is typically identical or very similar to GNSS signals.
In the early days of development, pseudolites also used the same frequency as GNSS, creating the promise of a seamless positioning technology – a single receiver that can be used outdoors and indoors. This dream was dashed when the signal frequency of Pseudolite devices was moved to the industrial, scientific and medical (band due to the risk of interference to GNSS signals.
But for all its promise, Pseudolite is relatively complex and expensive, and is likely to largely proliferate in applications where utmost accuracy is paramount.
A more cost-effective, high-accuracy indoor positioning technology is Ultra-Wideband (UWB). Ultra-Wideband is a communications technology that employs a wide bandwidth (defined as greater than 20 percent of the centre frequency, 500MHz). 500MHz bandwidth results in pulses of 0.16ns width, which means that it is possible to do accurate range-based positioning indoors, with a range measurement accuracy in the order of centimetres.
Only a handful of companies in the world: Ubisense, Timedomain, Zebra and Decawave, can provide robust UWB positioning systems. There are two significant issues with UWB-based positioning: synchronisation of “anchors” (or base stations) and non-line-of-sight (NLOS) errors.
The most reliable way to address the synchronisation challenge a wired connection between the anchors — making UWB infrastructure deployment a non-trivial task. A wireless method to achieve synchronisation is preferred, but presents its own challenges. Decawave claims that they can offer a wireless solution, however, the details have not yet been revealed.
To avoid the need to synchronise anchors, a technique for measuring the round-trip of the radio frequency signal has also been employed. Non-line-of-sight signals significantly increase the range measurement error, and there is no reliable way to solve this problem if the signal cannot penetrate the obstacles. Our own testing at UNSW has shown that UWB systems can achieve up to 10-20 centimetre positioning accuracy.
The CSIRO has developed their system, known as Wireless Ad-hoc System for Positioning (WASP), which also uses radio frequency signals, but with a narrower bandwidth than UWB.
WASP technology has been commercialised by Minetec to accurately locate and track mine workers in deep underground environments, and employed in tracking athlete data with wearables – allowing precise tracking and data collection within concrete velodromes and ice hockey rinks.
One method to bind them all?
As you may be beginning to understand, there is a dizzying myriad of sensors and technologies available for indoor positioning. We
know that fusing several sensors can normally deliver better positioning results, and the combination of several technologies can deliver higher availability.
This all suggests one key proposition – What if we could use all available sources for positioning?
So-called All-Source Positioning and Navigation (ASPN) has developed into a hot research topic. ASPN is different from sensor
fusion; it provides an all-in-one solution. It uses a uniform architecture to achieve data fusion of numerous sensors and data filtering, with the ability to re-configure. It aims to provide maximum accuracy and reliability with whichever navigation and positioning methods are available.
ASPN may be the ultimate – but it is not a simple solution. Everybody wants one solution for all scenarios, but so far there are no magic bullets.
There always looms a distant chimera, however. The most future positioning system sounds as cool as the technology is – the quantum positioning system.
Accelerometers based on super-cooled atoms could track one’s position with stunning precision.
Currently under development for defence applications in the US UK and Australia, testing is focusing on military submarine positioning, that currently rely on a relatively inaccurate system of accelerometer-based positioning while submerged. A submarine that goes a day without a GNSS fix will experience a navigation drift in the order of an entire kilometre –a distance that a quantum accelerometer could reduce that to a single metre. This technology is still far from mature; however, it could be a revolutionary positioning technology to replace all existing competitors.
Mix’n’match: technologies and use cases
Until such a mirage materialises into a usable, mature technology, the answer lies in perfecting a mix of technologies and techniques. We’ve found that to harness this vast array of techniques and technologies in a way that delivers a usable system, you need to develop a system that puts your users needs first – select techniques and technologies whose strengths play to these, and whose drawbacks are minimised for that case.
Our indoor positioning research team at UNSW has worked with a broad range of technologies, from WiFi to inertial measurement positioning (IMU); from fingerprinting to time-of-arrival techniques. We have tailored our indoor positioning technologies and systems to specific use cases, including as navigational aids for people who are blind and visually impaired (BVI) and tracking underground workers.
The SIMO project is our navigation and positioning system for BVI people, and our aims are simple. We intend to develop a usable, low cost positioning system for the visually impaired have been developing our solution for the last decade. Supported by the Australian Research Council (ARC), our partner Vision Australia and a startup called RSID, we started with a WiFi fingerprinting system, integrated with accelerometer, gyroscope and s as mobile stations.
Because of the difficulties in deploying WiFi access points and iOS’s security keeping WiFi closed to external positioning services, we moved to a Bluetooth Low Energy (BLE) network. We developed our own beacons for these networks, and tested our system at two sites: OfficeWorks North Sydney and UTS library, recording an average of two metres’ accuracy with BVI volunteers.
We made some important discoveries. Fingerprinting technology can achieve the best performance in complicated indoor environments – the obstacles make the fingerprints unique. However, for relatively open indoor environments, fingerprinting runs into problems without these distinguishing characteristics.
In order to give the visually impaired more freedom in most public areas, we decided to include Ultra Wideband positioning (UWB) in our system. The next phase of our system will incorporate a combination of Bluetooth and UWB to provide positioning accuracy of ranges between 10-20 centimetres to one or two metres, depending on the requirement of the specific location. have kindly joined as support organizations for this next phase.
Dr. Binghao Li is a VC’s Post-Doctoral Research Fellow in the School of Surveying and Spatial Information Systems, The University of New South Wales, Sydney, Australia. Binghao obtained B.Sc. in Electrical & Mechanical Eng. from Northern Jiaotong University, P.R. China in 1994 and M.Sc. in Civil Eng., Tsinghua University, P.R. China in 2001. He received his Ph.D. from the University of New South Wales, Sydney, Australia in 2006. His research area is pedestrian navigation, new positioning technologies and network-RTK.