The automatic vehicle monitoring to improve the urban public transport management
Luigi La Franca · 2004
ABSTRACT: In order to improve efficiency of the public transport system, the technological evolution has allowed the employment of innovative solutions for the automation of remote control systems of the bus fleet on the road network. These systems, said “AVM ” (Automatic Vehicle Monitoring) and introduced in the urban transport network, allow to offer the maximum levels of availability of the lanes, to improve the quality of the service (commercial speed, regularity of the transport, etc) and the information to the user. The recorded data about the vehicular fleet position can be also used for off-line optimising of the urban public transport routes and frequencies. In this work the recorded data about the vehicular fleet position of the public transport firm of Palermo (the main town of Sicily) have been elaborated using neural networks in order to calibrate a multidimensional function among independent variables and the value of commercial speed in a medium working day and for different traffic conditions (rush hour or non congested conditions) in different links of the transport network. A sensitivity analysis has been also carried out in order to determinate the importance of each variable in the commercial speed estimation. The calibrated model has been tested on different links (unseen from the neural network in the calibration process) of the public transport network of Palermo producing good results in the comparison between estimated