Fig 1. Among the protocol stack layers for IEEE 802.15.4, only layers 1 and 2 belong to the standard, while the upper layers are defined by the specific protocol (e.g., ZigBee, 6LoWPAN, and WirelessHART).
Fig 2. The 2.45-GHz PHY direct sequence spread-spectrum (DSSS) format is the most common for wireless sensor and actuator networks.
Fig 3. Nodes in a beacon-enabled frame structure are synchronized to the superframe boundaries by beacon frames broadcast at a regular interval from a network controller.
Fig 4. A cluster-tree network is used with 802.15.4 radios to create a wireless sensor network.
Since its 2003 release, the IEEE 802.15.4 radio has become hugely popular for wireless sensor and actuator networks (WSANs), and it is the underlying radio standard for protocols such as ZigBee, 6LoWPAN, and WirelessHART. This growth can be attributed to its availability as an inexpensive standard IC. It’s also sophisticated yet reasonably easy to use.
The IEEE 802.15.4 radio is a short-range, multichannel, spread-spectrum radio designed to support large, low-power, low-data-rate mesh networks. The standard and ZigBee are often, and improperly, considered synonyms of one another. Therefore, it’s important to understand the clear separation of functions implemented in the radio from those managed in the upper protocol layers (Fig. 1).
The 802.15.14 standard only specifies the lowest two layers of the protocol stack—the physical layer (PHY) and medium access control (MAC) layer. The upper layers of the protocol are separately defined by various interests according to their specific application needs. Of these, the largest and best known is the ZigBee Alliance. However, myriad proprietary protocols are also in use today, and they all rely on 802.15.4 as their underlying radio standard.
The current version of the IEEE standard is 802.15.4-2006. Three supplements (a, c, and d) have subsequently added more PHY options to the base specification. Most of today’s ICs comply with the base 802.15.4-2006 or compatible -2003 versions, which are the primary focus of the balance of this article. Free downloads of the specification and all supplements are available from the IEEE Standards Association Web site at www.standards.ieee.org/getieee802.
Spectrum Allocations
As with most standard product RF devices, the 802.15.4 radio uses carrier frequencies in the unlicensed UHF industrial, scientific, and medical (ISM) bands. ISM band allocations vary in different regions of the globe, so three frequency bands are supported: 868 to 868.6 MHz for use in Europe; 915 ±13 MHz for North America; and 2450 ±50 MHz, which is unlicensed in most regions of the globe.
The previously noted amendments to the base specification incorporate additional coverage for unlicensed sub-gigahertz bands used in Japan and China and Ultra-Wideband (UWB) operation in the 3.1- to 10.6-GHz band. In the U.S., the Federal Communications Commission (FCC) governs use of the ISM bands according to CFR Title 47 parts 18 and 15, which allow unrestricted access to the 915- and 2450-MHz bands within specified limits.
Although special licensing isn’t required, products using these unlicensed bands still must certify compliance with the FCC rules, predominantly with regard to output power and spectral content. With unrestricted access to these frequency bands, the ISM bands have become very popular, particularly the globally valid 2.45-GHz band that’s also utilized by Bluetooth and IEEE 802.11 radios, among others. Coexistence with these other services was carefully considered when designing the IEEE Std 802.15.4 radio.
Modulation Options
The 802.15.4 radio’s 868-MHz band is limited to a single data channel, while the 915-MHz band provides 10, and the 2.45-GHz band provides 16 (see the table). A maximum data rate of 250 kbits/s per channel is achievable in any of the three frequency bands, depending on the PHY mode. Four PHYs are specified in IEEE Std 802.15.4-2006:1
• An 868/915-MHz direct sequence spread-spectrum (DSSS) PHY employing binary phase-shift-keying (BPSK) modulation
• An 868/915-MHz DSSS PHY employing offset quadrature phase-shift-keying (O-QPSK) modulation
• An 868/915-MHz parallel sequence spread-spectrum (PSSS) PHY employing BPSK and amplitude-shift-keying (ASK) modulation
• A 2450-MHz DSSS PHY employing O-QPSK modulation
The 2.45-GHz DSSS PHY is by far the most common in WSAN applications. In this mode, the payload data are broken into 4-bit symbols, which are replaced by a corresponding 32-bit pseudorandom “chip” sequence (Fig. 2). This conversion has the effect of multiplying the original information bits by pseudorandom noise of a much higher frequency.
The 32-chip sequence is then demultiplexed, whereby even-indexed chips are sequentially fed to the QPSK modulator’s I-channel and odd-indexed chips are fed to the Q-channel with a half-cycle timing offset. Consequently, the modulator phase shifts by 1/2π every complex chip period TC, producing a 2M-chip/s rate in the RF channel.
Since the bandwidth occupied by each complex chip is inversely proportional to TC, the transmitted data’s RF energy is spread across a bandwidth that’s roughly 32 times what it would have been if it would have been directly modulated at the symbol rate. This lowers the power spectral density of the RF transmission and reduces the potential for narrowband interference.
The transmitted signal, which resembles white noise, is de-spread at the receiver using the same pseudorandom chip sequence to recover the payload data. Signals that aren’t modulated with the same pseudorandom codes are non-coherent noise to the receiver, giving a higher signal-to-noise ratio to the coherent code-modulated data of interest.
Potential Range
One of the most important design considerations for sensor networks—and often the first question that arises—is with regard to the effective link range between nodes. The 802.15.4 standard was developed with the general goal of operating within a typical link range of 10 meters, but doesn’t restrict applications of much longer distances. It minimally requires compliant devices to have a minimum transmit power level of –3 dBm and a minimum receive sensitivity of –85 dBm for the 2.45-GHz band or –92 dBm for the sub-gigahertz bands (where dBm = 10log(PT/1 mW).
Using the Friis transmission formula, the ratio between a transmitter’s output power and receiver’s sensitivity can be used to provide a first-order, idealized approximation of the potential range:
where PT is the power presented at the transmitting antenna in watts, PR is the power received by the receiving antenna, GT and GR are the gains of the transmitting and receiving antennas, λis the wavelength, and R is the distance in meters. Solving for R and substituting λ = 3 × 108/f:
With a 2.45-GHz 802.15.4-compliant radio meeting the minimum power specs and using unity-gain antennas, this gives a range approximation of:
The Friis model prediction assumes an unobstructed line of sight in free space. Such ideal conditions rarely occur in practical wireless sensor applications. True effective range is complicated by various other factors that are often determined empirically. These include attenuation and multipath fading from obstacles like walls or trees, local interference from other RF sources, the types of antennas used and their orientation, and elevation above ground or water.
One vendor’s 802.15.4-based street-lighting reference design supports a 1000-node network with node link ranges up to 1 km. Higher output power and better receiver sensitivity (a larger link budget) will extend effective range. However, higher transmit power levels will also adversely affect battery life. Fortunately, most 802.15.4-compliant devices offer a range of register-selectable transmit power levels, up to about +4 dBm, which provides the flexibility to adjust according to the application.
Access Method
The IEEE 802.15.4 radio uses a carrier sense multiple access with collision avoidance (CSMA/CA) algorithm to mitigate clashing transmissions in the shared RF medium. Because the radios are half duplex, they can’t detect collisions while transmitting, as in the manner used by wired Ethernet.
Instead, prior to transmitting, a device must first perform a clear channel assessment (CCA) to determine if the channel is idle. If idle, it transmits an entire packet of up to 128 bytes. If the channel is busy, transmission is deferred by a randomly generated delay and then retried. Reliable data transfers are ensured by using short acknowledgement frames.
If the packet’s acknowledgment request bit is set, an acknowledgment frame confirms successful reception. The absence of an acknowledgment frame within the timeout limit will cause the sender to retry the transmission. After a set number of retries, the upper layers of the protocol are notified of the communication failure.
To manage channel access, the standard supports both contention-based and contention-free modes. Contention-based access, also called the non-beacon-enabled method, uses unslotted CSMA-CA. With this method, a node can access the channel in a more ad-hoc fashion using the basic CSMA-CA algorithm.
The contention-free mode, also called the beacon-enabled method, uses slotted CSMA-CA to impose more structured access within a superframe that optionally includes up to seven guaranteed time slots (GTSs). Channel access is exclusive for devices transmitting in an assigned GTS, so the CSMA-CA algorithm isn’t used. Non-GTS devices still employ contention-based access, but are restricted to a contention access period (CAP) built into the superframe structure (Fig. 3).
Nodes in a beacon-enabled network are synchronized to the superframe boundaries by beacon frames broadcast at a regular interval from a network coordinator. In WSANs, this mode allows battery-powered sensor nodes to manage power more efficiently by synchronizing their sleep/active duty cycle to scheduled GTSs. It also provides a mechanism for the upper protocol layers to manage timing for devices that require low latency.
Network Configuration
Three operating roles are defined for nodes in an 802.15.4 network: a network device, a coordinator, and a personal-area-network (PAN) coordinator that acts as the master coordinator for the network. A network has only a single PAN coordinator, and any coordinator node can become the PAN coordinator during the formation of a new network.
Coordinator and network device nodes join a network by associating to a PAN coordinator. Coordinator nodes have the full MAC layer services to support peer-to-peer communication, and the simpler network device is typically a termination point.
The standard also distinguishes between a full-function device (FFD) and reduced-function device (RFD). An FFD can operate in any of the three defined roles, while an RFD only operates as a network device. FFDs communicate with many other FFDs or RFDs, but RFDs only associate with a single coordinator node (which must be an FFD) in a star network topology.
Coordinator nodes are responsible for managing the list of associated network devices. Furthermore, in the beacon-enabled mode, these nodes are responsible for broadcasting beacon frames and managing GTS scheduling. The coordinator’s peer-to-peer communication capability allows for the formation of more complex topologies, such as the cluster-tree (Fig. 4) or mesh network. Higher-level protocols such as ZigBee exploit this capability to create self-organizing, self-healing sensor networks.
Radio Availability
Several IC suppliers offer standalone IEEE 802.15.4 radios for under $3 in low volumes. They’re increasingly being provided as part of an integrated system-on-a-chip (SoC) that also includes a low-power microcontroller and memory, which is ideal for size- and power-constrained wireless sensor applications. SoC suppliers generally offer full software stacks for widely adopted protocols like ZigBee and 6LoWPAN, as well as simplified proprietary protocols that rely on the 802.15.4 radio.
For many, the quickest and most cost-effective path to market will be with a pre-certified wireless module. Modules ease the burden of RF layout design and circumvent the need for additional FCC certification testing. For 2.45-GHz-band devices, the small 1/4-wave dimension of 1.2 in. means the antenna can often be embedded right on the module. For more information about available SoCs and modules based on the 802.15.4 standard, go to Avnet’s Web site at www.em.avnet.com/smartnetworks.
Obviously, many other options exist for RF-based sensor networking, including a seemingly unlimited variety of proprietary devices and protocols. But the sophisticated, yet inexpensive, IEEE 802.15.4 radio and its adoption as the underlying radio standard for a wide array of protocols likely ensures its ubiquitous role in WSAN applications. Future adaptations of the standard, such as the precision ranging capabilities enabled by the Ultra-Wideband PHY, hint at even more interesting future applications.
Reference
1. IEEE Std 802.15.4-2006, “IEEE Standard for Information technology—Telecommunications and information exchange between systems—Local and metropolitan area networks—Specific requirements—part 15.4: Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (WPANs),” 2006.
2. Gutiérrez, Jose, Callaway, Edgar, and Barrett, Raymond, Low-Rate Wireless Personal Area Networks: Enabling Wireless Sensors with IEEE 802.15.4, Second Edition, IEEE Press, 2007.
3. Elahi, Ata, and Gschwender, Adam, ZigBee Wireless Sensor and Control Network, Prentice Hall, 2009.