Vehicle Lane Tracking with Vision/GPS/IMU



The GPS and Vehicle Dynamics Lab at Auburn University is currently developing methods to combine GPS/INS with computer vision/image processing techniques for highway lane tracking. Many companies in the Intelligent Transportation Systems community are developing methods for lane departure warning (LDW). The responsibility of an LDW system is to alert the driver of an un-intended vehicle lane departure. Currently, LDW systems use a vision system mounted on the vehicle to detect lane markers. These vision systems are environment sensitive and therefore are not always reliable.

The GAVLab is currently exploring methods to use GPS and inertial sensors to support the vision-based LDW systems. These methods will demonstrate how these additional sensors, which now come standard in many automobile lines, can be utilized for highway lane tracking. Increased lane tracking performance will lead to safer and more reliable advanced driver assistance systems (ADAS). Nissan has provided the GPS and Vehicle Dynamics Lab with an Infiniti M45 test bed for this experiment. The M45 is equipped with a commercially available technology package that includes an adaptive cruise control system (ACC) and a vision system for lane departure warning (LDW). The GAVLab has added data acquisition instrumentation to the car. The instrumentation includes an on-board PC for data collection and display, a differential GPS receiver, a typical commercial/automotive grade GPS receiver and controller area network (CAN) hardware to receive vision and inertial sensor data from the vehicle.

Figure 1:


Estimation of vehicle lateral offset between vehicle and lane marker using GPS-INS-Vision and map