Improving Position-Time Trajectory Accuracy in Vehicle Stop-and-Go Scenarios by Using a Mobile Robot as a Testbed

This study sets an example of how mobile robotic vehicles can be used effectively in research on intelligent transportation systems. Especially the stop-and-go mobility seen in heavy traffic conditions was simulated with a mobile robot, and the study is focused on how to obtain distance-time trajectories more accurately under these conditions. System identification tests of the mobile robotic platform, whose kinematic model was developed, were also carried out, and all solutions regarding robot movement were obtained. For the congested traffic simulation, various stop-and-go points are designated on a predetermined straight route segment to mimic behavior of a vehicle in congested traffic. Robot trajectories were obtained under different scenarios by using both GPS data and a kinematic model through the utilization of motor encoders. More accurate and consistent trajectories were achieved by fusing these trajectories with the Extended Kalman Filter. The main contribution of this study is demonstrating how the number of stop-and-go positions can improve the accuracy in estimating the robot/vehicle trajectory. The paper shows how the cumulative error in predicting the trajectories in reduced as the number of stops increases. For example, the trajectory estimated for a scenario involving five stop-and-go points is 94% more accurate than that for the case with a single stop. DOI: 10.61416/ceai.v25i3.8365