A preliminary study of respirable suspended particulate level in taxi transport interchanges in Hong Kong

A study on pollutant dispersion and distribution inside public taxi transfer interchanges (TTIs)is reported. The pollutant levels inside TTIs are affected by many factors, for example, taxidata, climatic conditions, human activities and geometrical layout of TTIs. A sitemeasurement of respirable suspended particulate (RSP) level is carried out in a typical TTI inHong Kong. After analysing the effect of the above factors on RSP level, we propose to useartificial neural networks (ANNs) to study such phenomena. The recorded data withindifferent time periods inside the selected TTIs are used as the test data set to train theproposed neural network model. The recovery performance of the ANN model is analysedand justified. In this study, we compare the forecasting results and the measured data of RSPin morning and afternoon sessions, respectively. The results show the feasibility andreliability of the proposed approach for forecasting pollutant levels inside TTIs.