Applications of Intelligent Transportation System (ITS) often require wireless communication networks to support navigation and location identification, which is emerging as one of important research issues in ITS. Among the various infrastructures for supporting ITS development, wireless sensor network is the most commonly used medium to transmit data from a target source to the receiving sensors, and/or between sensors. Such data, often referred to as geolocation data, can then be used to calculate the Time Difference of Arrival (TDOA) to the various sensors for the purpose of precise location identification. However, the limitation on the system bandwidth and energy resources motivates the use of data compression within the network, when data are transmitted between sensors. This paper is intended to present three data compression approaches specially designed for the geolocation data transmitted between sensors within the wireless communication network for the purpose of transportation target identification/locationfinding, including wavelet transform with arithmetic encoding method, wavelet packet analysis with Fisher-Information (FI) method, and TSU model (monotone increase method). The effectiveness of each presented method is evaluated by calculating the precision of TDOA from the decompressed data, which is utilized for determination of the location of the source. Sensitivity analysis is conducted on all the presented methods in terms of compression ratio (CR), Signal-to-Noise ratio (SNR), and computational time. A comparison among the three methods is provided in order to recommend the optimal compression tool for various applications.
Abstract