Across the U.S., the cost and efficacy of emission inspection and maintenance programs in place both at the state and country level has witnessed some criticism. In response, a Ph.D student at the Engineering and Public Policy and an associate at Civil and Environmental Engineering teamed up with a senior associate at Engineering and Public Policy.
The team developed a method to detect over-emitting vehicles using remote data transmission and machine learning. The new method would be less expensive and more effective than current inspection and maintenance programs in place.
Using the current program, most states in America make use of passenger vehicles to undertake periodic emission inspection and maintenance in a bid to ensure air quality. This is executed to ensure a vehicle’s exhaust emissions are within limit of the standards set at the time of manufacture of the vehicle.
On-board Diagnostics System currently monitors Emissions
Meanwhile, the onboard diagnostics systems of the car itself processes all of the vehicle’s data, including metrics through which emissions are monitored nowadays – a knowledge not known to all.
Effectively, the emission tests check if a vehicle’s check engine light is on. The system gauged over-emitting vehicles which was likely to be 87 percent true, and 50 percent times false passing of over-emitters when compared with tailpipe testing of actual emissions.
As cars increasingly become integrated into the Internet of Things akin to smart devices, this will require state and country administrations to force drivers to take their vehicles for regular inspection and maintenance checkups when all the required data is stored in the vehicle’s on-board diagnostics system.
In a bid to eliminate the unnecessary costs and improve the effectiveness of inspection and maintenance programs, the team published the study in IEEE Transactions on Intelligent Transportation Systems.