Smoothing Methods to Minimize Impact of Global Positioning System Random Error on Travel Distance, Speed, and Acceleration Profile Estimates

Abstract

The Georgia Institute of Technology is evaluating the feasibility and effectiveness of mileage-based pricing programs as transportation control measures. Incentives were provided to study participants who change driving behavior in response to cent per mile pricing (fixed pricing and pricing as a function of congestion level). In-vehicle Global Positioning System (GPS) devices were used to estimate distance traveled and driver behavior (e.g., speed and acceleration profiles). The accuracy of estimated mileage accrual speeds by road classification, and acceleration rates used in pricing algorithms, is paramount. Various data-smoothing techniques were applied to the instrumented vehicle GPS speed data, and performance of the algorithms was evaluated in minimizing the impact of GPS random error on speed, acceleration, and distance estimates. The conventional discrete Kalman filter algorithm was modified to enhance its ability to control GPS random errors. Each smoothing method produces different second-by-second speed and acceleration profiles ( t-test and χ2 tests) except for the Kalman filters. The techniques provided different travel distance estimates, but the modified Kalman filter was the most accurate compared with distance estimates from the onboard vehicle speed sensor monitor. The modified Kalman filter is the recommended technique for smoothing GPS data for use in pricing studies. Additional smoothing methods will be evaluated as they are identified.