Land-uses and activity-based traffic forecasting model

As road traffic in India is increasing rapidly, the accurate forecasting of traffic volume seems to be an important parameter for traffic control and transportation planning of an area. It seems to be more important in the case of rural roads traffic, as it contributes significantly to generating higher agricultural income and productive employment opportunities. A number of rural development policies, implemented by the Government of India, result in the increase of traffic day-to-day from rural areas. In general, the fundamental basis for the estimation of the total traffic volume of such areas is the land use pattern within and around it, the living standard or level defined by socio-economic factors pertaining to that area, population demographics, employment patterns, roads connectivity, etc. In this work, 30 villages from some districts of Uttrakhand and Wester Uttar Pradesh were selected as the study area. The data related to influencing variables that affect traffic generation was collected through random survey. Out of these 30 villages, data from 25 was used for the development of traffic prediction models, and the remaining was used for the validation of the developed model. Regression analysis and artificial neural network (ANN) were used to develop the models. Statistical measures were evaluated to examine the stability of the relationships developed and to select the best-fit models. These models can predict the traffic of a village at a given data of the significant influencing parameters and its growth rates. Finally, the models were compared by predicting the traffic volume of years 2011 and 2015 for the same study locations.