Improvements to a queue and delay estimation algorithm utilized in video imaging vehicle detection systems

摘要

Video Imaging Vehicle Detection Systems (VIVDS) are steadily becoming the dominantmethod for the detection of vehicles at a signalized traffic approach. This research isintended to investigate the improvement of a queue and delay estimation algorithm(QDA), specifically the queue detection of vehicles during the red phase of a signalcycle.A previous version of the QDA used a weighted average technique that weightedprevious estimates of queue length along with current measurements of queue length toproduce a current estimate of queue length. The implementation of this method requiredsome effort to calibrate, and produced a bias that inherently estimated queue lengthslower than baseline (actual) queue lengths. It was the researcher??????s goal to produce amethod of queue estimation during the red phase that minimized this bias, that requiredless calibration, yet produced an accurate estimate of queue length. This estimate ofqueue length was essential as many other calculations used by the QDA were dependentupon queue growth and length trends during red.The results of this research show that a linear regression method using previous queuemeasurements to establish a queue growth rate, plus the application of a Kalman Filterfor minimizing error and controlling queue growth produced the most accurate queueestimates from the new methods attempted. This method was shown to outperform theweighted average technique used by the previous QDA during the calibration tests. During the validation tests, the linear regression technique was again shown tooutperform the weighted average technique. This conclusion was supported by astatistical analysis of data and utilization of predicted vs. actual queue plots thatproduced desirable results supporting the accuracy of the linear regression method. Apredicted vs. actual queue plot indicated that the linear regression method and KalmanFilter was capable of describing 85 percent of the variance in observed queue length data.The researcher would recommend the implementation of the linear regression methodwith a Kalman Filter, because this method requires little calibration, while alsoproducing an adaptive queue estimation method that has proven to be accurate.