Signal setting in an isolated junction : before and after experimentation.

The application of different designed parameters in an isolated signalised junction using different methods was verified with before-and-after experimentation. The methodology is divided as follows: traffic flow and level of service indicators measurement in the present configuration for different days; design of signal parameters in relation to the data collected using different proposed models; estimation of forecasted values for traffic flow and user level of service indicators; application of signal setting parameters designed with different models during different time periods and different days; measurement of traffic flow and level of service indicators in the design configuration; comparison between forecasted and real traffic flow and level of service indicators for different days and different signal parameters. From the observation of user behaviour in the designed configuration some indications for best practice can be obtained; from the comparison between forecasted and real value indications about the objective forecasted in each model can be verified. The method was applied in an isolated junction situated in the central business district of Giarre (Italy) and is the most congested junction in the town. The junction has four arms and in the peak hour is used by 3000 vehicles/hour. Initially two cameras were installed near the junctions and on different days the flows divided into the different arms and different vehicle categories were counted. In relation to the counted flow and queues new signal setting parameters were designed using the Webster, Sigcap and Sigset methods. Forecast queues, saturations and user delays were estimated for each method. A set of parameters selected from that obtained from the designed methods, one for each time interval of one hour, were applied in the system during different days. In the new configuration the queues, saturations and user delays are measured on the system. The measured values are compared with those predicted. In the designed configuration, in the best case, there are improvements in the level of service at the junction and on average a reduction of about 30% in the user delay was obtained. Some indications are discussed concerning the comparison between forecasted values applying the three models and real values obtained by determining the range of validity of each model in the junction considered. Some considerations regard the saturation flow evaluated with the Highway Capacity Manual and real values obtained in the junction considered. From the observation of the junction in the different situations (present and designed with different methods) some best practice indications are extracted. For the covering abstract see ITRD E126595.