You may not realize it, but your car has thousands of sensors quietly gathering information about your driving.

An auto repair technician might tap into this trove of information by plugging a device into a port below the steering wheel and above the pedal to collect raw data on your vehicle’s health.

Using these ports to collect such data has been commonplace since the mid-1990s. But a small firm based in Columbia is attempting to change how that information is gathered and analyzed.

Agnik, a 20-person data-mining company founded by University of Maryland-Baltimore County professor Hillol Kargupta, recently patented technology to sort through a vehicle’s raw data to decide what should be sent to the cloud for analysis.

The technology cuts down on the amount of information to be transmitted, opening the door to all sorts of wireless applications that previously might have choked on the flood of unfiltered data.

“When you’re talking about these little embedded devices connected over wireless networks bringing on this huge amount of data, it’s not scalable,” Kargupta said.

Finding uses

His company’s solution, part of a field known as distributed data mining, is attracting attention from fleet operators, drivers and insurance companies interested in making new use of the data being collected by vehicles.

For instance, Agnik itself is selling a product called MineFleet, aimed at helping companies that own fleets of trucks and other vehicles keeps tabs on fuel costs and driver behavior.

Agnik’s Vyncs technology is in a consumer device called Car Connection that is sold at Sears for about $170 and manufactured by Audiovox. It allows drivers to monitor their car’s location, receive maintenance alerts and check other data.

Last year, Sprint and Agnik partnered to offer technology for insurance companies that would allow them to offer policies that base rates on a vehicle owner’s driving habits. The idea is to incentivize drivers to plug in a device monitoring their GPS, accelerometer and onboard diagnostics by offering discounts or rewards. Typically, both the driver and the company would be able to access a Web portal where they view a vehicle’s driving score and performance, Kargupta said.

It can be challenging for small players to enter the data mining market because of the “virtual oligopoly” of large firms such as Oracle, said Doug Kelly, industry analyst at IBISWorld. But the rise of cloud computing is changing the distribution and pricing models in the industry, he said.

“There are new market entrants introducing new technology, finding niche success, but they still have to contend with competition from the upper level,” Kelly said.

Beyond cars

Kargupta founded Agnik in 2004, while teaching in UMBC’s computer science and electrical engineering department — a job he still holds. Initially funded by small technology grants, Kargupta’s fledging data-mining company was formed out of an incubator at the university, where he hired two part-time employees. Within months of its founding, Agnik was attracting the interest of the Army, Defense Department and the Department of Homeland Security among others, all of whom were interested in Agnik’s method of machine-to-machine data transfer.

Though he’s been interested in the auto industry since he was approached by a Chrysler representative 17 years ago suggesting the application of data-mining methodology to cars, Kargupta said he hopes eventually to take the technology beyond cars into other industries. His broader interest is in machine-to-machine communication, which could eventually be used to develop a smart grid, or a more efficient electricity grid. But he’s also particularly excited about the possibilities for car data in social gaming or social media — allowing drivers to check in at certain locations, or integrate social media platforms with driver data.