Study On The Precision Improvement Of Infrared APC For Bus Passenger Counting And The Prediction Model Of Bus Passenger Transport Volume

The development of public transportation system (PTS) is the most efficient approach to resolve the traffic problems in a metropolis. Behindhand information technology in bus system has resulted in low efficient management and low bus service level currently. Therefore, information-based bus will help the bus management transfer from simple and intuitional pattern to scientific and intelligent one, and really improve bus service level.The mastery and prediction of bus passenger distribution information, plays a significant role in optimizing bus route net and the vehicle collocation of route, improving vehicle supervision and evaluating service level objectively. Aiming at the high error question of infrared APC, which is the most extensive application of passenger counting technology, this paper focuses on the method of improving passenger counting precision and the prediction model mend of bus passenger transportation volume. It is very important for improving bus passenger counting precision, implementing accurate prediction of bus passenger transportation volume and promoting informational and intelligent bus.This paper discusses the existing problems during the actual application of infrared APC in bus passenger counting, according to the analysis of three typical passenger statistic technologies: manual sample statistic technology, statistic technology of using IC card information and APC technology. It also studies the disposal technology of infrared APC data matching into route stations, the statistic error range of infrared APC, the reason of infrared APC error and the orderliness of infrared APC data, etc. Based on the regulations of infrared APC data, it puts forward a precision improvement technology about infrared APC counting buses passengers. Though the application of bus passengers collected practically and infrared APC data, the counting precision of infrared APC after improvement is improved clearly and the effectiveness of the model is validated.Based on the prediction model analysis of bus passengers transportation volume, this paper studies the regulations of passenger transportation volume data and discovers the property of changing periodically of bus passengers per mile aiming at improving prediction precision. Through analysis, it is proved that vehicle number and route mile are the key factors of influencing bus passenger transportation volume. Then this paper builds up a bus passenger transportation volume mend model based on the property of bus passengers per mile. The prediction model is validated through utilizing production data of a bus corporation in Chongqing city. The result shows that the maximal relative error of route passenger forecast data is under 15% and the average prediction error is half of the one by typical model. The bus passenger transportation volume prediction system completes the functions of transforming from chart into system database automatically, prediction result show and prediction precision show by chart or table.