More than a third of all traffic fatalities are due to skids and rollovers, which typically occur because a driver is moving faster than road conditions allow. In 2001, Fabio Romeo, then a senior manager at the Italian tire company, Pirelli, wondered whether new tire technology could help. He envisioned equipping cars with smart tires that would send out a warning when they were losing traction.
To realize this vision, he enlisted the help of his former Ph.D. advisor, Alberto Sangiovanni-Vincentelli
, who brought on board two of his Berkeley colleagues. The Berkeley researchers, together with Pirelli and several Bay Area companies, continue to push forward on the project with some of the most innovative tools and technologies Berkeley has to offer: energy-scavenging batteries, low-power radios, and new wireless protocols, all assembled using Sangiovanni's platform-based design approach (see Building Reliable Embedded Systems
). "There are so many unknowns in every direction, which makes this the ultimate engineering problem," Sangiovanni says. "We are pushing the envelope in every dimension.
The plan is to equip each tire with several sensor nodes that would use accelerometers to compute the deformation of the tire as it strikes the road's surface. A tire that is losing traction and spinning or sliding deforms less. The slippage information would be transmitted wirelessly to a microprocessor in the car, which would inform the driver or automatically make the appropriate adjustment.
The sensor nodes would consist of layers, "like a black forest chocolate cake," says Sangiovanni. The sensor would be one layer, while another would provide power, and another, the wireless communication capability. The sensors could send raw data, or one or more of the sensors could contain a microprocessor that would refine the data before sending it. A key factor in determining the right architecture, Sangiovanni says, is the amount of energy that will be available for the processors and the wireless connections.
Wireless sensor nodes are typically powered by batteries, but these sensor nodes would be buried under rubber and would need to be self-powering. Jan Rabaey
, who co-directs Berkeley's Wireless Research Center
, has been working with Paul Wright
, a professor of mechanical engineering specializing in energy-scavenging devices, to create an alternative mechanism.
Wright and Rabaey first built devices that extract vibrational energy from a tiny metal bar that bounces with the motion of the tire. That approach eventually evolved into a more efficient electromechanical device for converting the rotational kinetic energy of the tire into electricity.
To minimize power needs, Rabaey, a leader in creating cheap, low-power radios, is developing a wireless radio specifically for the project. "The requirements are not simple," says Rabaey, noting that the radio must be capable of beaming two megabits per second through water and snow while withstanding extreme forces. In addition, so as not to interfere with the function of the tire, the radio would have to weigh less than five grams and have an antenna that takes up less than a square centimeter. Currently, Rabaey's group is performing experiments with an old car scavenged from a junkyard.
Sangiovanni and his industrial collaborators, meanwhile, are devising appropriate protocols for the wireless connection, and Italian company Accent is helping to industrialize the design and ensure that the sensor system will be small, cheap, robust, and feasible to manufacture.
The researchers think they can handle these challenges to create a general prototype for this and similar applications. Still, Sangiovanni notes, tire manufacturers support their models for nine years, and there will need to be further progress on all fronts before the system components will be robust enough to work for that long in such a harsh environment. "You don't want to have to change your tires because the electronics inside them broke," he adds.