Abstract:
Application of fixed sensor and probe vehicle to collect traffic data has become popular over the years. Probe data provide spatial traffic information and direct measurements of travel time. However, their frequency is restricted with GPS reception errors in the probe data. On the other hand, video sensors record traffic data continuously at fixed locations only. So, the limitation of probe sensor in temporal domain and fixed senor in spatial domain tends to combine probe sensor with fixed sensor. Considering different characteristics of these traffic data from various sources, recently combination techniques are applied to reconstruct vehicle trajectories. This technique has become an advanced research area for intelligent transportation systems (ITS). Making an allowance for all-around limitations and advantages of different data sources, the combination technique is being developed to generate more reliable and continuous traffic information. Still, almost all the literature explore and improve over conventional lane based traffic condition. This may be due to the difficulty and high cost involved in data extraction, and complexities associated with non-lane based heterogeneous traffic movement. Addressing such shortage, this study deals with the development of a trajectory estimation method to reconstruct vehicle trajectories combining video sensor and probe data on urban arterial and freeway for non-lane based heterogeneous traffic condition. The combination technique involves extra equipment installation difficulties when the road segment becomes non-homogeneous due to traffic operations or lane configurations. Particularly, in developing countries, chaotic traffic pattern and absence of fixed sensor in the roadway create the trajectory reconstruction technique even more challenging. Video sensor (as well as other fixed sensors) based trajectory reconstruction technique requires sensors for each non-homogenous road segment and needs separate fundamental diagram (FD) to estimate required traffic parameters. This difficulty can be overcome through probe sensors alone. Thus, this research proposes an alternative approach of estimating traffic parameters only from probe data considering variable road capacity. Those parameters are used as input to reconstruct vehicle trajectories through lopsided network. The alternative method improves the Root mean square error (RMSE) of estimated travel time around 38.0% compared to that of original method where traffic parameters are obtained from FD, with respect to ground truth travel time. This method is found to be very appropriate, economical and reliable especially where required numbers of fixed sensors are unavailable.
