Speed-Merchant UWB Ushers In Wireless Video

Nov. 17, 2005
Despite its short range (less than 10 m), Ultra-Wideband and its up-to-1-Gbit/s data rate is a perfect fit for wireless streaming video.

Once again, we're on the verge of the launch of a new wireless technology that will change how we use our electronic products. And much like other wireless technologies, the change will be very positive.

With Ultra-Wideband (UWB), we will enjoy high-speed wireless connectivity in consumer products, especially TV and PCs. But like most other wireless technologies, UWB is taking more time than expected to get out the door. UWB is a wonky kind of wireless, but what wireless is easy?

But we're well past its early 2000s hype period. Now, companies are hunkered down doing the final work before product launch. Yes, there are a few products now, but the big push is around the corner. In fact, January's Consumer Electronics Show in Las Vegas promises to be the launching point for what will no doubt make 2006 the year of UWB.

Where does UWB fit in our wacky world of wireless? Take a look at Figure 1, which shows its niche. On the graph of range versus data rate, it achieves the highest data rate of any technology — but over the shortest distance. UWB easily produces a data rate of 100 Mbits/s up to 1 Gbit/s, yet range is typically less than 10 m.

It makes one wonder what such a strange technology's purpose is. But if you think about it, there is only one thing we haven't made wireless: video.

HOW IT WORKS UWB is a broadband wireless technology. Like spread-spectrum and orthogonal frequency-division multiplexing (OFDM), it spreads a signal over an incredibly wide bandwidth but at very low power. This offers four benefits.

First, broadband wireless technologies are better in applications that experience multipath propagation problems. The wider the bandwidth, the better the immunity to reflections and related propagation problems. Second, with wideband wireless, many signals can be placed on top of one another, creating a form of multiplexing. Third, wideband techniques produce signals that rarely interfere with other signals in the same spectrum. That's because their low power makes them appear more like noise than as interfering signals. Fourth, there's inherent security because it's so hard to detect and recover.

UWB has bounced around since the 1960s, when it was first discovered and developed in secret for secure military communications and radar. In its initial form, it used ultra-narrow baseband pulses to spread the signal over a huge bandwidth. Called impulse radio, it relied on the generation of a uniquely shaped pulse called a monopulse (Fig. 2). It generated a signal occupying a bandwidth that's greater than 20% of its center frequency, where the bandwidth is roughly the reciprocal of the pulse width. This characteristic usually forces UWB well into the microwave region.

In 2002, the Federal Communications Commission (FCC) allocated the spectrum from 3.1 to 10.6 GHz to UWB. This 7.5-GHz swath of bandwidth includes some of the most widely used wireless services, including Bluetooth and Wi-Fi in the 2.4- and 5.8-GHz bands, WiMAX, ZigBee, satellite radio, radar, some 3G cell phones, and heaven only knows what government and military wireless. Yet because the average transmit power level was set to an ultralow -41.3 dBm/Hz, virtually any UWB signal is going to look like low-level background noise to any other service.

Some initial concern was raised about potential electromagnetic-interference problems generated by UWB. But most experts now agree it's not a problem. Impulse UWB is generally called timemodulated or TM-UWB. Pulse position modulation (PPM) is the most common modulation method. The narrow pulses shaped with filters provide the wide bandwidth.

In direct-sequence (DS) UWB, the data to be transmitted is first modified by a unique higher-speed coded chipping signal, as are the signals in direct-sequence spread-spectrum (like cdma cell phones). This spreads the signal over a wide bandwidth and provides a way to "channelize" the bandwidth for the simultaneous transmission of many signals and their recovery. Modulation is either phase-shift keying (PSK) or PPM. DS-UWB transmitters are super simple and use very low power, but the receiver and its complex correlation recovery circuits are somewhat more of a challenge.

Today, several versions of TM/ DSUWB products are available. Examples are Artimi's PPM product, Freescale's Xtreme Spectrum Trinity PSK product, and Pulse-Link's CWave DS product. These work exceptionally well, consume ultra-low power, and easily achieve data rates in excess of 100 Mbits/s at a range of 10 m.

For some reason, though, the industry hasn't adopted these techniques as standards. In fact, most companies already have abandoned the impulse approach and are diving head-on into a new, more complex OFDM UWB standard: multiband OFDM (MB-OFDM). This technology will form the foundation for most of the coming UWB products.

MB-OFDM divides the UWB spectrum into multiple 528-MHz wide bands, used in groups of three. The lower three bands, ranging from 3.168 to 4.952 GHz, make up the initial spectrum to be used, mainly because it's relatively easy these days to make all-CMOS radio ICs in this space. The center frequencies for these three 528-MHz bands are shown in Figure 3.

In turn, a 128-band OFDM signal occupies each band. Each band has a width of 4.125 MHz. Of the 128 carriers, 100 are used for data, 12 are pilot carriers, and the remainder form guard bands. A bit stream using binary PSK or quarature PSK modulates each carrier, depending on the data rate. The resulting signal is then transmitted continuously in one of the three bands. Alternately, the data may be hopped from one band to another in one of four basic patterns.

The MB-OFDM radio uses the standard coding, scrambling, and inverse fast Fourier transform (IFFT) to generate the signal to be transmitted. At the receiver, an FFT recovers the original signal. Consequently, digital signal processing lies at the heart of an MBOFDM UWB radio. Nonetheless, a 128-point FFT isn't that complex and can be implemented with logic in a small space these days. The resulting radio can achieve a data rate of up to 480 Mbits/s at about 2 to 3 m and up to 110 Mbits/s at 10 m.

UWB APPLICATIONS Video streaming is the most often mentioned use of UWB. It makes sense because digital video needs lots of bandwidth, even in compressed format. Moreover, there happens to be a shortage of practical technologies that can handle video, especially in a portable or mobile application.

Overall, UWB could provide the desired cable replacement for applications that carry video. Or it could become the link to a video monitor or screen, especially those big wall-mounted plasmas and LCDs. Anything carrying digital video is a candidate for a UWB wireless connection. Even a video camera is a possibility.

Because UWB uses very low power, it's especially attractive for battery-powered devices like cameras, camcorders, MP3 players, and remote speakers.

Although video is most likely a key UWB application, emphasis has shifted to another cable-replacement scheme — wireless USB ports. Because the USB port has emerged as the easiest to use and most ubiquitous connection to PCs and laptops, it makes sense to make a wireless version that can carry data rates up to the maximum 480 Mbits/s.

Most UWB manufacturers are targeting the wireless USB as the killer application for UWB. As long as they make it entirely transparent to the user, it will find a home. But if they require special configuration procedures or new drivers, consumers will just get aggravated.

Another possible major application is the mesh network. Mesh networks rely on very small low-power, and in most cases, battery-powered nodes with transceivers designed primarily for wireless sensor networks. Most applications only need low speed.

So far, ZigBee/802.15.4-standard transceivers have captured the biggest share of the mesh market. Wi-Fi, or 802.11a/b/g, also works, but it's more expensive and eats more power. The benefits are longer range and higher data rates than ZigBee.

Mesh networks are designed so each node is a repeater to pass along data from one node to the next, greatly extending the range of any node. The problem is that data rates are severely limited with ZigBee and other technologies. That's where UWB comes in.

Its range is limited to about 10 m. Using multiple nodes in close proximity can extend the range many times over. Furthermore, data rates of over 100 Mbits/s or even higher can be sustained. Only a standard DS-UWB or MB-UWB chip set with the appropriate software is needed to make the ultimate fast mesh.

A potential application opportunity for UWB is to become the new physical layer (PHY) for Bluetooth. The Bluetooth SIG recently announced that it was considering UWB as a future higherspeed PHY for Bluetooth. Bluetooth's base data rate of 1 Mbit/s is fast enough for most audio applications. A recent new version called Enhanced Data Rate (EDR) provides a transmission speed of 3 Mbits/s. Yet to remain viable, Bluetooth must provide a path for higher speeds to add new applications profiles. While it's undecided as to which UWB technology will be used, all are being considered. UWB is a good fit no matter what, providing the higher speed needed over the short-range domain of Bluetooth.

It will be interesting to see what other-applications will emerge once the chips become available. Potential wireless applications include a wireless IEEE1394 interface or a wireless version of the High-Definition Multimedia Interface (HDMI). Many consumer electronic companies use both to transmit video and audio between products.

ISSUES WITH UWB For some time, the lack of a standard bogged down UWB's progress. Once the FCC blessed UWB in February 2002, standards work began in the IEEE wireless personal area networks (WPAN) group (802.15.3a). Many initial proposals were quickly boiled down to two key approaches: DS-UWB and MB-OFDM.

Over the years, much effort was expended to reach a consensus on one standard. It never happened. The MBOFDM group, which represents by far the greater number of companies, has never been able to outvote the Freescale/ Motorola block with its DS-UWB approach. With this stalemate, both groups went their own way.

Recently, the Multiband OFDM Alliance (MBOA) and the Wimedia Alliance merged and completed work on a standard their member companies agree on. Wimedia will manage the standard and provide Wi-Fi-like testing and certification to ensure interoperability of competing chip sets and products.

With so many companies supporting this standard, it's obvious which UWB approach will be used by most end products. But the DS-UWB approach has found its own niches, so it will no doubt coexist with Wimedia MB-UWB.

At this point, an IEEE standard seems dead in the water, but progress marches on. Is a standard that provides ubiquitous interoperability in the consumer arena really necessary?

Another issue affecting UWB is the overall state of the world markets for such products. For any wireless product to succeed, it must have a world market. The main markets for video and wireless USB outside the U.S. are in Europe, Japan, and Korea. No formalized regulations exist in any of those areas yet. They're expected to be finalized in 2006. They won't be the same as those for U.S. UWB, but they will be close enough that the products developed here will work as is or can be readily modified.

One issue that continues to bother some is competition from other wireless technologies, especially 802.11a/g and the forthcoming 11n standard. Wi-Fi already can transmit video because the upper data rate of 54 Mbits/s is plenty fast enough, and even its range is above UWB. The problems are complexity and power. The latter isn't a factor in most consumer products that plug into the ac line, though.

The 802.11n standard, still under negotiation in the IEEE Task Group N, will feature data rates from 100 to 250 Mbits/s and higher at a range of up to 50 to 100 m using multiple-input multiple-output (MIMO) antenna technology. could be a threat to UWB in ac-powered products.

A new group, the Enhanced Wireless Consortium (EWC), consists of 27 companies looking for a fast solution to the standard delay. The two main within the IEEE working on this TGn Sync and WwiSE. To date, still no consensus, and little progress has been made of late. The hopes to move the process forward to a final standard or take off on own, like the Widmedia group.

The standard put forth by EWC combines features of both standards proposals now being debated in the IEEE. It supports data rates to 600 Mbit/s using the 20-MHz assigned bands or 40-bands where the spectrum is available. It's compatible with 802.11a/g, and supports spatial multiplexing for simultaneous transmission using up to 4 antennas. With "big gun" Wi-Fi companies like Atheros, Broadcom, Cisco, DLink, Intel, and others, the pressure is on finalize 802.11n fast or the EWC will go its own way. There's no both technologies will vie for consumer video applications, and which will come out on top is still to be determined.

WHO'S WHO IN UWB Lots of companies are addressing the UWB opportunity. Now, most are ready to announce chip sets and related products. End products should be available for the first time in 2006. The table (p. 42) provides a quick overview of those participants, as well as the various organizations that support the UWB field.

See Figure 4


To join the conversation, and become an exclusive member of Electronic Design, create an account today!