Font Size: a A A

An Adjustable Time Scheme For Traffic Flow Aggregation In Wireless Sensor Network

Posted on:2012-06-12Degree:MasterType:Thesis
Country:ChinaCandidate:L YuanFull Text:PDF
GTID:2218330368488255Subject:Computer technology
Abstract/Summary:PDF Full Text Request
In recently years, the problem of the urban transport is really serious with the increasing of the social economic. Intelligent transportation system is an important way to alleviate traffic problems. Wireless sensor network are used to collecting traffic data by intelligent transportation system because of its high precision and low cost. With the increasing of the urban population, traffic acquisition system will need to collect more traffic data, otherwise, the heavy traffic challenge the store of the data. Traffic control system has a high real-time requirements for the traffic flow, the effective real-time traffic information is not only make the correct identification of the traffic status but also give the real-time control decisions. Wireless sensor network are composed of sensor nodes whose energy is limited, so increasing the life of the wireless sensor network also should be addressed.The traffic data is collected by a fixed time in the wireless sensor network, that is within a fixed time intervals, the traffic data is collected and aggregated and then transfer to the traffic control center. However, historical data show that traffic flow in different times will have great change, which will lead to great fluctuations for traffic flow, speed in very small time interval, such fluctuations may be used in real-time traffic control, so if we used fixed-time traffic data aggregation can not guarantee the transmission data is real-time and that is not conductive to increasing the life of the wireless sensor network.In this paper, we propose a variable time traffic flow aggregation mechanism combine with the state of the urban transport. By analyzing history of collected traffic flow parameters and analyze the binomial distribution of the traffic flow, we get the relationship between share of vehicles and aggregation time. Then we use neural network to forecast the flow and share of vehicles. At the end, we get the variable time of the aggregation data based on the forecast data. Thereby increasing the real-time data and effectively reduced data traffic across the network and extend the lift of the wireless sensor network.BP neural network is widely used in traffic flow prediction, however, it the training itself is slow and can not guarantee convergence to the global minimum. The genetic algorithm is a global optimization algorithm, so we use genetic algorithm to optimized the topology of the neural network, then we could get the more accurate forecast traffic flow and the more reasonable polymerization time. Simulation experiments show that the method improved real-time traffic data, and increase the life of the WSN network to a certain extent.
Keywords/Search Tags:wireless sensor network, aggregation time, traffic flow prediction, neural network, genetic algorithm
PDF Full Text Request
Related items