Developing GPS Antenna Error Models For Improved Centimeter Level Positioning
- Rhonda Franklin, Professor, Electrical and Computer Engineering
- Demoz Gebre-Egziabher, Assistant Professor, Aerospace Engineering & Mechanics
- Robert Sainati, Adjunct Professor, Electrical and Computer Engineering
A 2017 study of lane departure/keeping systems involving 25 states found that these systems reduced crash rates and injuries by 11% and 21%, respectively. Since 2000 these systems have been available but only on high-end vehicles. Key enablers for these systems are positioning systems with centimeter level accuracy. Typical GPS receivers used in vehicles today do not meet this requirement. Real Time Kinematic GPS (RTK GPS) systems with the requisite performance that do exist are expensive (> $10,000) and cumbersome. Thus, they are not suitable for automotive applications. Currently video, laser, radar or infrared sensors are used to determine where the car is relative to visible lane markings. Though less costly, their integration into a car can only be done when the automobile is manufactured since these sensor systems are an intimate part of the vehicle's highly integrated electrical system. If after-market systems that can provide an accurate position solution could be purchased for a reasonable cost, significant driver/road safety benefits could be achieved. Prior work by the authors of this proposal has shown that a key factor limiting the accuracy of RTK GPS systems is the quality of the antenna measurement. On accurate (and expensive) GPS receivers, the antennas are large, uniform and come with extensive factory calibration. On low-end receivers the antennas are small, variable from one unit to the next and lack extensive factory calibration. The net effect is inaccurate positioning whose accuracy depends on the direction-of-arrival of signals transmitted by the GPS satellites. Since a minimum of 10+ GPS satellites are normally in view, invariably GPS signals arrive from different directions, thus introducing time varying error. The overall objective of this work is to develop a methodology for characterizing this error and subsequently developing mathematical algorithms that can compensate for this error in RTK GPS receivers.
- Start date: 03/2018
- Project Status: Active
- Research Area: Transportation Safety and Traffic Flow