The Bluetooth Frame Structure

The Bluetooth Frame Structure

Bluetooth defines several frame formats, the most important of which is shown in two forms in Fig. 4-36. It begins with an access code that usually identifies the master so that slaves within radio range of two masters can tell which traffic is for them. Next comes a 54-bit header containing typical MAC sublayer fields. If the frame is sent at the basic rate, the data field comes next. It has up to 2744 bits for a five-slot transmission. For a single time slot, the format is the same except that the data field is 240 bits.

If the frame is sent at the enhanced rate, the data portion may have up to two or three times as many bits because each symbol carries 2 or 3 bits instead of 1

bit. These data are preceded by a guard field and a synchronization pattern that is used to switch to the faster data rate. That is, the access code and header are carried at the basic rate and only the data portion is carried at the faster rate. Enhanced-rate frames end with a short trailer.

Let us take a quick look at the common header. The Address field identifies which of the eight active devices the frame is intended for. The Type field identifies the frame type (ACL, SCO, poll, or null), the type of error correction used in the data field, and how many slots long the frame is. The Flow bit is asserted by a slave when its buffer is full and cannot receive any more data. This bit enables a primitive form of flow control. The Acknowledgement bit is used to piggyback an ACK onto a frame. The Sequence bit is used to number the frames to detect retransmissions. The protocol is stop-and-wait, so 1 bit is enough. Then comes the 8-bit header Checksum. The entire 18-bit header is repeated three times to form the 54-bit header shown in Fig. 4-36. On the receiving side, a simple circuit examines all three copies of each bit. If all three are the same, the bit is accepted. If not, the majority opinion wins. Thus, 54 bits of transmission capacity are used to send 10 bits of header. The reason is that to reliably send data in a noisy environment using cheap, low-powered (2.5 mW) devices with little computing capacity, a great deal of redundancy is needed.

Various formats are used for the data field for ACL and SCO frames. The basic-rate SCO frames are a simple example to study: the data field is always 240 bits. Three variants are defined, permitting 80, 160, or 240 bits of actual payload, with the rest being used for error correction. In the most reliable version (80-bit payload), the contents are just repeated three times, the same as the header.

We can work out the capacity with this frame as follows. Since the slave may use only the odd slots, it gets 800 slots/sec, just as the master does. With an 80-bit payload, the channel capacity from the slave is 64,000 bps as is the channel capacity from the master. This capacity is exactly enough for a single full-duplex PCM voice channel (which is why a hop rate of 1600 hops/sec was chosen). That is, despite a raw bandwidth of 1 Mbps, a single full-duplex uncompressed voice channel can completely saturate the piconet. The efficiency of 13% is the result of spending 41% of the capacity on settling time, 20% on headers, and 26% on repetition coding. This shortcoming highlights the value of the enhanced rates and frames of more than a single slot.

There is much more to be said about Bluetooth, but no more space to say it here. For the curious, the Bluetooth 4.0 specification contains all the details.

RFID

We have looked at MAC designs from LANs up to MANs and down to PANs. As a last example, we will study a category of low-end wireless devices that people may not recognize as forming a computer network: the RFID (Radio Frequency IDentification) tags and readers that we described in Sec. 1.5.4.

RFID technology takes many forms, used in smartcards, implants for pets, passports, library books, and more. The form that we will look at was developed in the quest for an EPC (Electronic Product Code) that started with the Auto-ID Center at the Massachusetts Institute of Technology in 1999. An EPC is a replacement for a barcode that can carry a larger amount of information and is electronically readable over distances up to 10 m, even when it is not visible. It is different technology than, for example, the RFID used in passports,which must be placed quite close to a reader to perform a transaction. The ability to communicate over a distance makes EPCs more relevant to our studies.

EPCglobal was formed in 2003 to commercialize the RFID technology developed by the Auto-ID Center. The effort got a boost in 2005 when Walmart required its top 100 suppliers to label all shipments with RFID tags. Widespread deployment has been hampered by the difficulty of competing with cheap printed barcodes, but new uses, such as in drivers licenses, are now growing. We will describe the second generation of this technology, which is informally called EPC Gen 2 (EPCglobal, 2008).

EPC Gen 2 Architecture

Often, the tags look like stickers that can be placed on, for example, pairs of jeans on the shelves in a store. Most of the sticker is taken up by an antenna that is printed onto it. A tiny dot in the middle is the RFID integrated circuit. Alternatively, the RFID tags can be integrated into an object, such as a driver’s license. In both cases, the tags have no battery and they must gather power from the radio transmissions of a nearby RFID reader to run. This kind of tag is called a ‘‘Class 1’’ tag to distinguish it from more capable tags that have batteries.

The readers are the intelligence in the system, analogous to base stations and access points in cellular and WiFi networks. Readers are much more powerful than tags. They have their own power sources, often have multiple antennas, and are in charge of when tags send and receive messages. As there will commonly be multiple tags within the reading range, the readers must solve the multiple access problem. There may be multiple readers that can contend with each other in the same area, too

The main job of the reader is to inventory the tags in the neighborhood, that is, to discover the identifiers of the nearby tags. The inventory is accomplished with the physical layer protocol and the tag-identification protocol that are outlined in the following sections.

EPC Gen 2 Physical Layer

The physical layer defines how bits are sent between the RFID reader and tags. Much of it uses methods for sending wireless signals that we have seen previously. In the U.S., transmissions are sent in the unlicensed 902–928 MHz ISM band. This band falls in the UHF (Ultra High Frequency) range, so the tags are referred to as UHF RFID tags. The reader performs frequency hopping at least every 400 msec to spread its signal across the channel, to limit interference and satisfy regulatory requirements. The reader and tags use forms of ASK (Amplitude Shift Keying) modulation that we described in Sec. 2.5.2 to encode bits. They take turns to send bits, so the link is half duplex.

There are two main differences from other physical layers that we have studied. The first is that the reader is always transmitting a signal, regardless of whether it is the reader or tag that is communicating. Naturally, the reader transmits a signal to send bits to tags. For the tags to send bits to the reader, the reader transmits a fixed carrier signal that carries no bits. The tags harvest this signal to get the power they need to run; otherwise, a tag would not be able to transmit in the first place. To send data, a tag changes whether it is reflecting the signal from the reader, like a radar signal bouncing off a target, or absorbing it.

This method is called backscatter. It differs from all the other wireless situations we have seen so far, in which the sender and receiver never both transmit at the same time. Backscatter is a low-energy way for the tag to create a weak signal of its own that shows up at the reader. For the reader to decode the incoming signal, it must filter out the outgoing signal that it is transmitting. Because the tag signal is weak, tags can only send bits to the reader an a low rate, and tags cannot receive or even sense transmissions from other tags.

The second difference is that very simple forms of modulation are used so that they can be implemented on a tag that runs on very little power and costs only a few cents to make. To send data to the tags, the reader uses two amplitude levels. Bits are determined to be either a 0 or a 1, depending on how long the reader waits before a low-power period. The tag measures the time between low-power periods and compares this time to a reference measured during a preamble. As shown in Fig. 4-38, 1s are longer than 0s.

Tag responses consist of the tag alternating its backscatter state at fixed intervals to create a series of pulses in the signal. Anywhere from one to eight pulse periods can be used to encode each 0 or 1, depending on the need for reliability. 1s have fewer transitions than 0s, as is shown with an example of two-pulse period coding in Fig. 4-38.

EPC Gen 2 Tag Identification Layer

To inventory the nearby tags, the reader needs to receive a message from each tag that gives the identifier for the tag. This situation is a multiple access problem for which the number of tags is unknown in the general case. The reader might broadcast a query to ask all tags to send their identifiers. However, tags that replied right away would then collide in much the same way as stations on a classic Ethernet.

We have seen many ways of tackling the multiple access problem in this chapter. The closest protocol for the current situation, in which the tags cannot hear each others’ transmissions, is slotted ALOHA, one of the earliest protocols we studied. This protocol is adapted for use in Gen 2 RFID.

The sequence of messages used to identify a tag is shown in Fig. 4-39. In the first slot (slot 0), the reader sends a Query message to start the process. Each QRepeat message advances to the next slot. The reader also tells the tags the range of slots over which to randomize transmissions. Using a range is necessary because the reader synchronizes tags when it starts the process; unlike stations on an Ethernet, tags do not wake up with a message at a time of their choosing.

Tags pick a random slot in which to reply. In Fig. 4-39, the tag replies in slot 2. However, tags do not send their identifiers when they first reply. Instead, a tag sends a short 16-bit random number in an RN16 message. If there is no collision, the reader receives this message and sends an ACK message of its own. At this stage, the tag has acquired the slot and sends its EPC identifier.

The reason for this exchange is that EPC identifiers are long, so collisions on these messages would be expensive. Instead, a short exchange is used to test whether the tag can safely use the slot to send its identifier. Once its identifier has been successfully transmitted, the tag temporarily stops responding to new Query messages so that all the remaining tags can be identified.

A key problem is for the reader to adjust the number of slots to avoid collisions, but without using so many slots that performance suffers. This adjustment is analogous to binary exponential backoff in Ethernet. If the reader sees too many slots with no responses or too many slots with collisions, it can send a QAdjust message to decrease or increase the range of slots over which the tags are responding.

The RFID reader can perform other operations on the tags. For example, it can select a subset of tags before running an inventory, allowing it to collect responses from, say, tagged jeans but not tagged shirts. The reader can also write data to tags as they are identified. This feature could be used to record the point of sale or other relevant information.

Tag Identification Message Formats

The format of the Query message is shown in Fig. 4-40 as an example of a reader-to-tag message. The message is compact because the downlink rates are limited, from 27 kbps up to 128 kbps. The Command field carries the code 1000 to identify the message as a Query.

The next flags, DR, M, and TR, determine the physical layer parameters for reader transmissions and tag responses. For example, the response rate may be set to between 5 kbps and 640 kbps. We will skip over the details of these flags.

Then come three fields, Sel, Session, and Target, that select the tags to respond. As well as the readers being able to select a subset of identifiers, the tags keep track of up to four concurrent sessions and whether they have been identified in those sessions. In this way, multiple readers can operate in overlapping coverage areas by using different sessions.

Next is the most important parameter for this command, Q. This field defines the range of slots over which tags will respond, from 0 to 2^Q−1. Finally, there is a CRC to protect the message fields. At 5 bits, it is shorter than most CRCs we have seen, but the Query message is much shorter than most packets too.

Tag-to-reader messages are simpler. Since the reader is in control, it knows what message to expect in response to each of its transmissions. The tag responses simply carry data, such as the EPC identifier.

Originally the tags were just for identification purposes. However, they have grown over time to resemble very small computers. Some research tags have sensors and are able to run small programs to gather and process data (Sample et al., 2008). One vision for this technology is the ‘‘Internet of things’’ that connects objects in the physical world to the Internet (Welbourne et al., 2009; and Gershenfeld et al., 2004).

Frequently Asked Questions