MODELING TRAVEL DEMAND AND CRASHES AT MACROSCOPIC AND MICROSCOPIC LEVELS
Analytics
40 views ◎21 downloads ⇓
Abstract
Accurate travel demand / Annual Average Daily Traffic (AADT) and crash predictions helps planners to plan, propose and prioritize infrastructure projects for future improvements. Existing methods are based on demographic characteristics, socio-economic characteristics, and on-network (includes traffic volume) characteristics. A few methods have considered land use characteristics but along with other predictor variables. A strong correlation exists between land use characteristics and these other predictor variables. None of the past research has attempted to directly evaluate the effect and influence of land use characteristics on travel demand/AADT and crashes at both area and link level. These land use characteristics may be easy to capture and may have better predictive capabilities than other variables. The primary focus of this research is to develop macroscopic and microscopic models to estimate travel demand and crashes with an emphasis on land use characteristics. The proposed methodology involves development of macroscopic (area level) and microscopic (link level) models by incorporating scientific principles, statistical and artificial intelligent techniques. The microscopic models help evaluate the link level performance, whereas the macroscopic models help evaluate the overall performance of an area. The method for developing macroscopic models differs from microscopic models. The areas of land use characteristics were considered in developing macroscopic models, whereas the principle of demographic gravitation is incorporated in developing microscopic models. Statistical and back-propagation neural network (BPNN) techniques are used in developing the models. The results obtained indicate that statistical and neural network models ensured significantly lower errors. Overall, the BPNN models yielded better results in estimating travel demand and crashes than any other approach considered in this research. The neural network approach can be particularly suitable for their better predictive capability, whereas the statistical models could be used for mathematical formulation or understanding the role of explanatory variables in estimating AADT. Results obtained also indicate that land use characteristics have better predictive capabilities than other variables considered in this research. The outcomes can be used in safety conscious planning, land use decisions, long range transportation plans, prioritization of projects (short term and long term), and, to proactively apply safety treatments.