CHOOSING THE OPTIMUM ARCHITECTURE FOR UWB RTLS

The choice of UWB RTLS architecture will impact the ability to achieve the greatest ROI from these new capabilities.

Key Considerations:

n Constant Tracking of Assets Everywhere

n Cost of Coverage and Installation

n Cost of Ownership

Some system architectures are designed to only provide locating in specific areas of coverage. This is a result of architecture choices that make the system too expensive for continuous coverage. Typically these architectures use discrete, multi sensor cells with inward focused antennas for each area of coverage, with each cell being a complete subsystem. For the purposes of this paper, we’ll term this discrete cell architecture. Unfortunately these architectures can add cost, complexity, and the potential for error associated with cell handoff

Discrete cell architecture can affect not only the cost of coverage, but the cost of installation. Systems built with discrete cell architecture must rely on sophisticated radio channel schemes, which increase the cost of site design and increase the complexity of post installation tuning. Although there can be some benefits, these same results can be achieved by other means without adding the complexity of additional long range radio links

Cost of Ownership can also be affected by architectural fundamentals. The addition of two way radio communication using non-UWB channels draws significantly higher power from the tag and can thwart one of the prime advantages of UWB: Ultra long battery life at very fast update rates.

By choosing a system architecture that is designed to provide complete coverage over large areas and utilizes simple beacon UWB radio links, assets or people can be constantly tracked throughout an entire enterprise in a cost effective manner, over a long period of time, without interruption.

Additionally, it is possible to have tags that blink at a rate of once per second, but will still blink continuously for 7 years without the complexity of motion sensors and other sources of field failures. Since the update of rate of these tags is so high, a second radio to manipulate that update rate is not required. This simplifies installation, virtually eliminates post installation tweaking, and minimizes maintenance.

Introduction

Ultra Wideband RTLS can maximize its ROI when it is able to provide its primary function: accurate tracking of resources (assets and people), in real time throughout the factory, store, office building, or even sports arena. 

One of the key reasons to invest in an RTLS system vs. an RFID system is to replace sporadic portal level information with a continuous and persistent flow of location information. 

The value of RTLS is realized by knowing the exact location of any resource, at any moment in time, and taking immediate action based on that knowledge. As an example; if a system is tracking tools, knowledge of the exact location of those tools when they are in the wrong place can be as important as the location data provided when matching them to a target process and work cell.

UWB RTLS has unique attributes that make it specifically qualified to provide this continuous stream of information. These unique attributes are:

n Precision Locating

n Fast Update Rates

n Long Battery Life

n High Capacity

Architectural Challenges

Omni-directional antennas look in all directions evenly. Directional antennas focus in a particular area to extend the range in that direction. In order to achieve this increased range in the intended direction, these antennas lose range in every other direction as shown in the Figure 1. If a system were designed to only cover a small area, then a pragmatic approach to maximize coverage would be to require all the antennas to be directional and focus them inward as shown in Figure 2.

AoA Locating

n Position calculated from the angles of arrival of a single signal at a minimum of two sensors

n Accurate estimation of angle of arrival is difficult (small inexpensive hardware can be limiting)

n Angle is heavily affected by signal reflections 

n Location confidence lowers significantly as distance from sensor increases

If a system is designed with an architecture that uses the inward looking discrete cellular architecture shown above, then it is possible to design a sensor that houses several antennas that can be used in an array to determine the Angle of Arrival (AoA) of the incoming signal. AoA determines the direction that the signal is coming from and uses the input from two or more sensors to determine a point of transmission as shownin Figure 5. Requiring only two sensors could be an advantage.

However, some researchers have found that these systems provide a low confidence of accuracy1.This low confidence is caused by reflections which maybe mistaken for true Line Of Sight signals and therefore introduce error in the angular measure of the received signal. Additionally, any inaccuracy in the determined angular measurement is amplified as the distance between the sensor and the tag increases. This error is in the shape of a cone that starts at the sensor and expands with distance as shown in Figure 7.

GPS provides amazing accuracy given that the satellites used for determining location are in orbits that are more than 20,000 km above the earth’s surface. This accuracy is achievable by using time based locating. 

Similarly the most accurate RTLS systems are also time based. The most common technique is Time Difference of Arrival (TDOA) which compares the exact arrival time at different reception points as shown in Figure 6.

Therefore, measuring the exact time that a signal arrives is critical to developing an accurate system. UWB does not place data on a sinusoidal waveform as is done with all conventional radio systems. Instead of modifying a sinusoidal wave by changing its amplitude,frequency, or phase, UWB systems carry data in the form of bursts (pulses) of energy. The fact that the signal is carried in an energy bursts across a very wide frequency band makes it easier to find the exact leading edge of the signal with sub-nanosecond accuracy.

These very wide-band signals make it possible to determine the first Time of Arrival, (TOA) of a signal, and ignore reflections. This is what makes time based UWB RTLS so accurate.

1. A transmitter (tag) sends a signal that is received by location sensors in (at least) three different locations.

2. Each sensor notes the signal’s time-of-arrival (TOA).

3. The difference in arrival times at any pair of sensors implies that the transmitter was located somewhere on a calculated hyperbola.

4. Using two sensor pairs (one sensor may be common between these pairs) implies that the tag resided at the intersection of two different hyperbolas.

A system designer that has chosen to use a discrete cell architecture capable of providing AoA location may choose to combine AoA with TDOA in order to achieve high precision RTLS with only two sensors. Unfortunately, while this does improve the likelihood of a better locate, it is not equivalent to TDOA using three sensors. As shown in Figure 7, while the time based TDOA hyperbola is quite accurate, the actual location could still be anywhere on the hyperbola within the intersection of the two cones of uncertainty. Given that the discrete cell architecture requires so many additional sensors at the intersections of the cells, this is a poor trade-off from a cost and therefore ROI perspective.

Optimal Architectural Choices

The first of these is to use systems designed with new generation, simple one-way beacon architecture.

This saves tag cost, and greatly reduces downtime and maintenance costs by virtually eliminating battery changes. The complexity of having a second traditional radio can be eliminated with very small losses in function by adding a simple low frequency Exciter function. 

An exciter is a low frequency electromagnetic field (EMF) device with very short, well-defined area of coverage, for which receivers can be made which consume extremely low power. Exciters do not require data connections and have been used in RTLS systems for over 15 years. The receiver side in the tag operates on micro-amps of current and does not have a significant impact on battery life. This technology can be used to create a transactional event in areas where there is only presence detection, and they can also be used to reprogram blink rates.

The choice of a continuous architecture utilizing multiple antenna types eliminates cellular handoff delays and saves significant cost if complete areas are covered. The entire facility can be covered and tags will not be out of sight between cells.

Although utilizing single point omni-directional antennas to provide continuous coverage will not support Angle of Arrival AoA locating, this is an easy tradeoff to make given the low confidence level of AoA locates. In fact, the large savings in infrastructure density from the 1-to-4 sensor tradeoff can easily allow additional sensors in suspected trouble spots such as areas with mobile or multiple blockers. This will result in locates throughout the facility with minimum variation in accuracy.

Conclusions

Simple beacon architecture based on international standards with continuous areas of coverage provides the

best solution for a high reliability, cost effective RTLS solution. It provides the four key requirements:

n Continuous and Constant Locating Coverage

n Increased Interoperability Through Use of Standardized Hardware

n Better Price Performance Through Reduced Cost of Coverage and Installation

n Lowest Long Term Cost of Ownership