Designers get a handle on wrist communicator
By Peter Clarke
EE Times
(10/27/99, 6:27 p.m. EDT)
Comic-strip readers and engineers have dreamed about the wrist communicator ever since Dick Tracy started using one in his battles with evil-doers. Today the villains challenging a Philips design team working on just such a gizmo are hardware and software architectures, power sources and manufacturing issues.
The results of Philips' compact personal communicator research program could, in the next century, become as commonplace as the mobile phone is today. The team predicts its communicator will be a viable commercial product in three or four years. The appliance will handle a set of applications that could range from data and video communications to personal health monitoring, audio and video entertainment, finance and shopping.
The task of filling in the details of what has, so far, been a science fiction concept has been given to a small group of researchers led by Miriam Vriens within Philips Semiconductors' Systems Laboratory in Eindhoven, Netherlands. Although many companies have shown one-off demonstrations of Dick Tracy-style technology, all the various versions have been limited in one or more critical ways, and none has been commercially viable, said Vriens.
It's no Movado, but team leader Miriam Vriens' watch is a classic melding of form and function: She can make and receive calls, with caller ID and a silent vibrating alarm. The Philips Systems Lab's wholistic approach to architectural decisions is consistent with system-on-chip design.
"Our job is to look at the idea from a systems point of view, in terms of the very complex ICs and software that will be required, and by looking at all the different production technologies," she said.
Already, Vriens can demonstrate the concept. Although the wristwatch is not the most stylish ever seen, and the "mobile phone terminal" is nothing more than an open pc board, she can make and receive calls on the gadget, complete with caller ID and a silent vibrating alarm to tell her when calls are coming in.
The Systems Laboratory is a think tank that works for Theo Claasen, chief technology officer at Philips Semiconductors. It is separate from the corporate R&D function, and works on more targeted development. It includes about 500 people at sites in Eindhoven; Hamburg, Germany; Southampton, England; and Taipei, Taiwan.
A project like the compact personal communicator has a number of benefits for Philips, Vriens said. By looking at the whole system, the group gets to make architectural decisions in software and hardware that, in the era of system-on-chip integration, are key to determining form, function and power consumption. It also means the group can determine where limitations lie in the ancillary technologies, such as displays, packaging, connectors and batteries. The current thinking is that the device will be based on a two-piece architecture connected by a Bluetooth short-haul radio interface and powered by one or perhaps several RISC processor cores from ARM Ltd.
"It's also a technology driver. Independent of the success of the project, we can still get to use the technologies developed [for the communicator] in other areas," said Vriens, adding that ultralow-power semiconductors and packaging are two that are likely to have widespread significance. "It also promotes broad cooperation between all the elements of Philips, in areas of passives, ceramics, antenna design," she said. "It's very beneficial."
Vriens said that powering the watch-based interface is one of the challenges her team faces. Can it be done with kinetic energy, by body warmth — or will it take a lithium ion battery?
"Kinetic power is not enough. For kinetically operated watches you only need microwatts, but an RF interface will require milliwatts," Vriens said. But new, flexibly formed lithium ion batteries could make it possible to house the battery in the wrist strap.
"We've calculated we could store 100 mA-hr in the wrist strap, which might be enough for two to six months' use," she said. "That assumes about 10 minutes use per day."
Similarly, the team is debating whether the Bluetooth antenna should be in the watch strap or in the watch.
To achieve the long-term aims for the personal communicator project, Vriens has two short-term goals to be realized over the next one to two years: create a technology demonstrator, and identify a co-development partner. That partner is likely to be Philips Semiconductors' lead customer for its compact personal communicator platform.
The unit will feature Scampi, the Simple Communication Accessory for Mobile Phone Interfacing, which is defined as a means to bridge the telecommunications and watch worlds. This approach was taken in part because the likes of Ericsson, Nokia, Motorola, Swatch, Seiko, Timex and Casio have the sort of sales volumes and money to provide a good hunting ground for a technology partner.
The compact personal communicator program began in June 1998, with three engineers from Eindhoven and two in Hamburg. Over the next four to five years the group is expected to define the platform architecture for the hardware and software building blocks that will enable the future design of such systems.
Already several key architectural decisions have been made. For example, the project will leverage existing technologies and use a Bluetooth radio-frequency interface to link a light, highly compact but functionally rich wristwatch interface to a more conventional mobile phone. The phone could stay in the wearer's pocket but provide the cellular communications link to the outside world.
For the local interface to the pc-board mobile terminal in the working model, Bluetooth is used. From the pc board out it uses a Digital Enhanced Cordless Telephone (DECT) connection, although that could equally well be a second-generation cellular telephone interface such as GSM or CDMA. In the not too distant future a third-generation air interface, such as wideband CDMA, could be used.
"Bluetooth is the most appropriate solution," said Vriens. "It is low cost, less than $5, and low power at about 1 mW, which helps address health issues. Bluetooth will be a standard feature on mobile phones, with an installed base of more than 100 million phones by 2002," she predicted.
The platform is currently based on the ARM processor architecture. "ARM is a standard in telecom and is very low power," Vriens said. "The ARM7TDMI core is used in the DECT interface and in the Bluetooth interface." For speech recognition functions Philips has a dedicated chip set called Spheric. This allows recognition of key words and phrases and allows calls to be made and answered hands-free.
ARM family cores are already specified within Philips Semiconductors' telecommunications terminal platform, so its selection was a natural for Vriens. But what happens if the ARM RISC architecture is eclipsed by another processor in terms of performance or power efficiency, parameters that are crucial to Scampi?
"We expect to see ARM respond to the demands of the industry," she said. "We could change to another core but there would have to be a commanding reason. Bluetooth is based on ARM so we would find it hard to move independent of that."
For Vriens and Scampi the next stage is to start integrating. By mid-2000 she expects to have squeezed the terminal into just two chips plus memories: a CMOS baseband chip and an RF chip in Philips' Qubic BiCMOS process technology. By mid-2001 she expects to have the functionality on a single CMOS chip, plus an extended set of applications.
|
