Online-offline Interactive Urban Crowd Flow Prediction toward IoT-based Smart City

Abstract : Urban crowd flow prediction is very challenging for public management and planning in smart city applications. Existing work mostly focuses on spatial and temporal dependence based flow prediction that are not well suited for predictions of instantaneous flow change usually due to social emergency incidents and accidents. In this paper, we propose an Online to Offline Interaction based Dilated Casual Convolutional Neural Network framework (O2O-DCNN) for crowd flow prediction. Both online attention behavior and offline crowd shift factors are considered, in case to capture the dependence between them and make more accurate predictions especially for instantaneous flow variations. The online and offline features are processed by dilated casual convolutions and then put into CBOW model based full connected network to make interactions. The performance evaluations are based on realistic User Detail Record (UDR) dataset of a southern city in China provided by China Unicom. O2O-DCNN is compared with the other related baselines in terms of MASE and MAE. The results show that our framework is with much better accuracy, especially for instantaneous flow variation scenario.

 EXISTING SYSTEM :

 ? The system would be made self-sustainable by developing services on top of the existing smart city infrastructure. ? The efficient management of large amount of data gathered by wireless sensor networks (WSNs) is a major challenge due to restricted capabilities of sensor nodes in terms of memory, energy, computation and scalability. ? A Sensor cloud can provide a promising solution to the above mentioned problem by combining a powerful and scalable massive storage infrastructure with the sensor networks for real-time processing and storing the WSN data as well as their analyzes. ? It is used to interconnect the end devices to the main communication infrastructure of the city.

 DISADVANTAGE :

 ? The main issue regarding to system scalability can be the high concentration of people in the areas where the system is deployed. ? The “Participatory Sensing” app provided a mean for the citizens to send information about the city status, including cultural events or problems (e.g. broken bin, dirtiness, etc.). ? Attending to the Area Under the Curve (AUC) parameter, we can see that all the classifiers have a very good performance for this classification problem. ? Radio Frequency systems provide highly accurate results in the short range and their use does not introduce any major privacy issue. ? When the decision mechanisms are applied offline, the size of the analyzed dataset is not a major issue.

 PROPOSED SYSTEM :

 • In, the researchers have pointed out the major obstacles on the way of smart city market and henceforth have proposed a procedure to bootstrap the smart city business based on the concept of big data exploitation. • The proposed approach in defines three stages to facilitate sustainable smart city development. • Several new start-ups that provide innovative services and deliver new applications and solutions for different smart city areas have been proposed in. • Some control servers are deployed within the proposed city network infrastructure to store the huge collection of data in its internal database.

 ADVANTAGE :

 ? This poorer performance of the LRC at the airport is mainly due to the location of the deployed devices that causes higher correlations. ? The three mechanisms analysed have shown a good performance in deciding whether the targets were inside or outside the building. ? It is important to highlight that they were chosen because of their lightweight computational footprint and fast convergence time. ? We intend to include other classification techniques, including more computational demanding algorithms, to evaluate which provides the best performance for positioning tasks. ? Therefore, we have studied the product-moment correlation coefficient from the measurements.