A brief overview of RFID technology and testing challenges for complex operating environments where throughput and communication issues may be caused by multiple readers, dense-mode environments and pre-existing non-RFID signals
The applications of radio-frequency identification (RFID) are growing rapidly as equipment prices drop and global markets expand. The use of embedded RFID is increasing. Coordinating bodies such as the Ubiquitous ID Center and the T-Engine Forum have been formed. And the GSM Association now supports embedding of RFID-based near-field communications capability into cellular phones.
However, a big challenge in RFID is optimising throughput (data reading speed) in complex, or even harsh, RF environments. Passive RFID tags may respond to any reader or readers in range. Protocols exist to work with this behaviour, but the result is a complex communications behaviour that can be difficult to test without the right equipment. In addition, embedded RFID systems may need to function, when integrated into the same device, with cellular, WLAN, Bluetooth or Zigbee technologies. Finally, interference from other users in the same band needs to be considered.
The result is a need to simulate complex RF environments and analyse the performance of the RFID system under these conditions before deployment. The pulsed nature of RFID and the typical interference sources make this task all the more difficult.
RFID technology overview
An RFID system at its simplest consists of a tag, which may be passive, and a reader. Architecturally, reading passive tags is somewhat different from the traditional full-duplex data link. Unlike traditional active data links, passive tags rely on the received RF energy to power themselves. These also do not generate their own transmit carrier signal. Rather, they modulate some of the energy being transmitted by the interrogator to the tag in a process known as backscattering.
By changing the loading of the tag’s antenna from absorptive to reflective, a continuous-wave (CW) signal from the interrogator can be modulated. This process is very similar to using a mirror and the sun to signal someone at a distance. It also eliminates the need for precision frequency sources and power-hungry transmitters in the tag. Since the reader and the tag share the same frequency, these must take turns sending information. Backscattering restricts communications between the reader and the tag to a half-duplex system.
Since the uplink from tag T to reader R (denoted by T→R) is modulated from the interrogator’s CW signal, it is possible to use spread-spectrum techniques such as frequency hopping. Any spreading or hopping on the interrogator’s signal is automatically removed in the homodyne down-conversion of the receiver, as it shares the same local oscillator signal.
This simple system turns complex when multiple tags, multiple readers and interference are present. Let’s take a look at two RFID design challenges that come from these deployment issues.
Multiple-reader and dense-mode environments
The broadband nature of passive RFID tag presents some challenges for dense (multiple) reader sites. Since the tag reader sets the system’s frequency of operation, and the tag is a broadband device that responds to any reader, the tag has limited ability to respond to a specific reader. Passive tags may try to respond to all readers that are interrogating them.
Many RFID systems are implemented in a multiple-reader or dense-mode environment. Here are some definitions:
Single-reader environment: A single reader is operating in an environment.
Multiple-reader environment: The number of simultaneously operating readers is less than the available number of channels.
Dense-reader mode: It’s the most challenging environment, where the number of readers is greater than the number of channels.
Reader and tag interference can take place within the operating environment—the zone within which the reader’s RF signal is attenuated by less than 90 dBc (a radius of approximately 1 km in free space). Consequently, many readers end up operating in a dense-mode environment whether by design or due to neighbouring RFID readers.
A warehouse application with fixed readers and accurate spectrum planning is likely to have minimal interference from neighbours within 1 km. However, a mobile RFID device should expect a dense-reader mode environment due the lack of control over safe mitigation distances. In this case, it becomes critical to discover what signals may be present in the environment where the RFID system will be, or is, deployed and understand the behaviour of the reader and tags in the presence of interference.
To handle this environment, ISO18000-6C readers that have been certified for dense environments often switch to Miller-modulated sub-carrier (MMS) encoding. This elaborate encoding provides more transitions per bit and is therefore easier to decode in the presence of noise. But it is slower for the same tag’s backscatter-link frequency (BLF).
Three different MMS schemes are available: Miller-2, Miller-4 and Miller-8. The number specifies how many BLF periods define a data symbol. For example, using the slowest BLF of 40 kHz, the data rate for Miller-8 is BLF/8 = 5 kbps. At such a slow rate, transmitting a 96-bit EPC and 16-bit error check will take 22.4 ms, corresponding to less than 45 tag reads per second (even fewer when all the overhead, such as the forward link commands, is included). Transmitting at this slow rate is not desirable for throughput reasons, but also because some regulations (The United States FCCs’ Part 15, for instance) allow operation at a single frequency only for an average of 400 ms within a 10- or 20-second period depending on the 20dB bandwidth of the signal. This regulation may require the tag reader to vacate the channel after 400 ms and hop to a different frequency even if not finished with reading at that frequency.
Readers and tags operating in conformance to ISO18000-7 take a different approach. These use longer RF transmissions with slower transfer rates, which allows the signal to be more immune to interference. For shipping container applications using the commercial equivalent version ISO 18185, this requires that the maximum transmission duration be increased to 60 seconds while maintaining a 10-second minimum silent period between transmissions (FCC part 15.240). At such slow transfer rates, transferring the full 128 kilobytes of data needed to identify all the contents of a shipping container can take up to two minutes. The tags used in accordance to this standard are active; they have an on-board power supply and tend to radiate at higher power levels than passive tags.
Both of these techniques imply that the testing solution needs to collect detailed RF data on pulsed signals over a relatively long time period.
Testing solutions for a dense-mode environment
Simulating a dense-mode environment is possible using an arbitrary waveform generator (AWG). Modern AWGs can be programmed to directly generate RFID signals across HF and UHF bands and thus be used to simulate a variety of signals such as multiple reader or multiple tag signals with just one instrument. This reduces the time and cost of having to configure multiple signal generators.
The analysis device often needs quite deep memory to capture all of these lengthy interactions. Typically, the tag reader tries multiple queries and may command the tags to reduce their link frequency to verify that the tags are vacating the channel as required in some implementations. Real-time spectrum analysers (RTSAs) are capable of analysing these types of transactions.
An RTSA can directly verify the 60-second transmission and the 10-second silent period of ISO18000-7, with memory depth of more than 100 seconds in this application. This allows a thorough analysis of error conditions.
It is also possible to use multiple acquisitions to analyse hopping and bursting RFID signals. In this mode, the RTSA is set up to capture data for a user-defined period of time whenever a hop, and the associated trigger, occurs. This, in combination with a very high frame rate (over 48,000 fps), allows capture, analysis and demodulation of a hopping RFID signal.
Once the signal is captured, it can be analysed in a way that helps engineers to understand whether the reader and tag are performing as expected in the current RF environment, and if not, why. Measurements of bit time, CW time, and response time between the reader and the tag (known as turnaround time) give important insight into the reader and tag interaction and throughput. Checking amplitude glitches against frequency events can help spot root causes of errors. If a particular bit is not decoding properly, is it because of an error in the FSK or ASK modulation? Correlating data in various domains helps to answer these types of questions.
Latest RTSAs can correlate data in frequency, time, symbol and other domains, allowing extensive and quick analysis of complicated RF environments and physical-layer interactions. For ISO18000-6C (EPC GEN2) signals, which automatically change their data rate, these instruments can automatically detect the symbol rate and highlight the preamble, making analysis even easier.
Monitoring RFID co-channel interference
RFID transceivers must comply with local regulations regarding creating interference as well as be designed for optimal immunity to interference. Allocated spectrum, for instance, varies from 2 MHz in Singapore and Europe to 26 MHz in North America. This makes different modulation schemes and collision avoidance techniques favoured in different parts of the world.
Two approaches taken to implement collision avoidance and reduce self-interference are frequency hopping and listen-before-talk (LBT)/synchronisation of RFID readers. Frequency hopping is utilised in the United States according to FCC 47 CFG Ch. 1 Part 15. LBT or synchronisation is implemented in most European countries according to ETSI EN 302 208-1.
These signals are bursty by nature in an environment subject to bursty interference from multiple readers, multiple tag responses, and even other RF services such as Wi-Fi, Zigbee, Bluetooth and similar short-range RF communications. Effective analysis of RFID signals in a realistic environment can be a complex task.
One of the best available monitoring techniques is RTSA digital phosphor capability known as DPX. This technique uses a very fast frame rate in combination with indicating signal density, or dwell time, by colour. This allows a unique view of pulsed RF signals in a complex environment.
Fig. 6 shows a simulated complex RF environment, created by placing a number of tags within read range of a reader. After only 30 seconds of monitoring the frequency-hopping output of the reader, a lot of information can be seen.
Red signals are always present. In this case, that means the noise floor and several interfering signals towards the bottom of the display. Green signals (mostly bursty interference in this example) are present perhaps 50 per cent of the time, and blue signals are infrequent, as indicated by the signal density scale in the lower right corner.
The blue signals are mostly RFID signals for communications between a reader and the set of tags. In this case, the modulation type is amplitude-shift keying, and the higher narrow blue pulses are 1’s and the lower narrow blue pulses are 0’s. DPX provides visibility of signals that would not be seen on a conventional swept analyser.
In this screen capture, the reader was successful on frequencies without much interference. This is shown first, by seeing that the mostly blue RFID pulses only occur on clear frequencies. Second, by looking at the other colours on the mostly blue RFID pulses, the extended dwell time of a successful RFID transaction can be spotted. It’s realistic to conclude that successful polling occurs at frequencies where no interfering signal is present and the signal-to-noise ratio is greater. This clearly shows that the chances of a successful tag read increase in an environment with minimal interference.
When doing frequency planning to restrict each reader to a certain channel (or channels), DPX can be used to ensure that the modulation sidebands are not at such a level that would cause interference in those channels being used by co-located readers. Note that the reader and tag signals in the centre of Fig. 6 have wide spectral spreading and are dwelling for a longer period of time than in other channels. The brighter signal skirts indicate a higher signal density and thus a longer dwell time. This could be a source of read failures in the adjacent channels, and action should be taken to ensure that filtering in the readers is sufficient to have immunity to this interference.
To sum up
RFID signals, by their nature, may be subject to complex, and even harsh, RF environments. In addition, the pulsed nature of RFID signals makes analysis difficult with conventional spectrum analysers. AWGs and RTSAs make effective simulation and analysis of multiple-reader, dense-mode environments and common interference signals practical. This technique can be used to ensure reliable RFID communications and throughput in harsh conditions.