Linux on Wheels: A New Opportunity

by Linley Gwennap

Linux fans may be aware of the Empeg in-dash MP3 player, a Linux-powered computer that stores and plays back gigabytes of MP3 files in the car. (See “MP3 Linux Players” by Craig Knudsen, LJ, July 1999.) The Empeg player is an undeniably cool device, but it is just the tip of the proverbial iceberg. Within five years, almost every car will have a powerful computer that provides a variety of services. Linux is a contender to win a large share of this market, which has no dominant player and will generate tens of millions of sales per year.

Modern automobiles are already a seething mass of computing resources, packing anywhere from 30 to 80 networked microprocessors that control windows, doors, airbags, lights, gauges, brakes, engine timing, temperature and anything else that moves or changes. These inexpensive chips are too small and slow to run Linux. But now automobile equipment makers are planning to take the next step, turning the car into a “$30,000 wireless Internet device,” says Mike Iannitti, director of Intel's In-Car Computing Division, with only a hint of a smile.

A Compelling Feature Set

Several concepts are converging to make an in-dash computer appealing. Various single-function devices exemplify these concepts today. The Empeg player, for instance, brings to the car the advantages of MP3, including instant access to a huge number of songs. Other digital music formats, such as digital FM radio and satellite radio, will become popular in the next few years, all requiring some sort of decompression software. A programmable platform will be flexible enough to handle new algorithms and protocols. Within five years, any car without these capabilities will be as out of date as a car with an AM radio and an 8-track player is today.

Navigation assistance is a second useful feature. Some high-end vehicles and rental cars come equipped with GPS systems that determine the car's location and plot a route to a specified destination, often providing audible directions (“turn left in 100 feet”). Anyone who has ever gotten lost while driving will appreciate the value of such a system. And unlike backseat drivers, it can be turned off.

A third area is safety and security. General Motors already has 150,000 customers for its OnStar system, which uses the GPS and a cell phone to contact a call center if the vehicle crashes or is stolen, providing the car's exact location. The call center can also unlock the car's doors if you forget your keys, and can help with other problems. Future systems could improve security by validating a voiceprint or fingerprint before allowing the engine to start.

Another helpful function is information access. This doesn't mean running a browser on the dashboard; rather, it is an intelligent mechanism for retrieving and delivering relevant information such as traffic reports, restaurant or ATM locations, or even just e-mail messages. The OnStar call center can direct a driver to a nearby hotel or restaurant. Other services already provide wireless Internet access.

Voice recognition will be a key feature in these car computers, as people can't type or press buttons at the same time they're trying to drive. By the end of this year, Delphi will begin shipping a Palm V docking station that allows drivers to access address and datebook entries using voice commands. The unit has a built-in cell phone that can be “dialed” using voice or by selecting an address book entry. Each of these functions is appealing to certain drivers. But the combination of these functions, particularly if the price is right, will be unstoppable. Acceptance of in-dash computers will be driven by two factors: platform standards and price.

Rapid Adoption Path

Today's first-generation products typically cost $800 to $1,000 and perform only one or two of the functions listed above. Current prices are bloated, because upfront development costs must be spread across a small number of units. As production volumes rise from tens of thousands to tens of millions, prices will drop dramatically. The cost of key components, such as microprocessors, flash memory and GPS receivers, is also falling due to normal semiconductor trends. For example, a GPS chipset that cost $100 two years ago sells for about $30 today, and will drop to $10 in another two years. These trends will help bring the price of an in-dash computer down to a few hundred dollars within the next few years.

The value of an in-dash computer will increase as vendors realize that all of the important functions require a fast microprocessor, a flexible operating system and a wireless interface. Combining digital music, navigation, information access, security and voice recognition into a single device will add little hardware cost but significant end-user value. The bigger difficulty is in creating the necessary software, particularly since new technologies and services are likely to continue to be developed after the car is purchased. Digital music formats are in flux and not likely to settle down for at least a few years. Information services will continue to evolve, and the broad deployment of car computers will undoubtedly spur a new round of innovation.

Ideally, the car computer should be built around a standard platform that allows third parties to develop applications. Without a standard, consumers will be confused and application development stunted. Microsoft hopes to play this role with Windows CE, which is used in Delphi's Palm V dock, Clarion's AutoPC and Visteon's ICES among others. But Linux's low cost, simplicity and open-source model should be attractive to auto equipment makers. WinCE faced little competition for these early systems, but more recently, a throng of embedded Linux vendors has risen to the challenge. The stakes are high. Worldwide auto sales totaled 54 million last year, about half the size of the desktop PC market. Full-function in-dash computers will appear as standard equipment in some model-year 2002 cars (just over a year from now) and will move rapidly through the product lines. Ford, for example, boldly claims that all its cars will have Internet capability within three years. This hot new market gives Linux the opportunity to establish itself as the platform of choice.

email: linleyg@linleygroup.com

Linley Gwennap (linleyg@linleygroup.com) is the founder and principal analyst of The Linley Group (http://www.linleygroup.com) analysis firm in Mt. View, California. He is a former editor in chief of Microprocessor Report.

Load Disqus comments