| With the rapid development and construction of the rail transport, it becomes faster and more efficient, which brings higher request both of the security and the intelligence for the rail transport. How to keep the security of the rail transport under the condition of high-speed and high-efficience is a key problem. Meanwhile, the key point of the safety run of the train is to acquire the running state of the train and the environmental contion exactly, which is intelligent processed by the train control system. The application of the Internet of Things(IoT) on the rail transport is an efficient way to deal with the key point. The position estimation of the train and the sensed informantion has an important role in the both factors,which leads the automation localization become a key problem in intellengent rail transport and receive considerable attentions Wireless sensor network (WSN) and radio frequency identification (RFID) are the two key technologies in the IoT, which also play important roles on the train's localization and the environment monitoring. In this dissertation, we will study the localization based on RFID and WSN respectively, including the design of the localization schemes, study and analysis the reliability and security of the localization schemes in complicated environment.Firstly, we present a simple and effective RFID localization method with good accuracy using passive UHF RFID tags for the tag localization. To achieve the accurate location, a path loss model is proposed to study the power of the reader as a function of read range. The function is analyzed based on the Friis transmition function and the empirical model that often used in indoor environments in theory and demonstrated in experiments. Based on the function, a circle tag localization method using multiple RFID antennas is proposed. The two and three antennas scheme are simulated and experiments are carried out to demonstrate the performance of the localization method. Experiment result show that the two antennas scheme can achieve an average localization error of 57cm, while the three antennas scheme decreases the localization error to 18cm with 90% of the localization error less than 25cm.Both experiment result and simulation result show that the proposed scheme can achieve good performance in the indoor localization. However, it's needed to note that the UHF RFID radio is easily affected by the indoor environment such as multipath indoor.Secondly, due to the feature of UWB great capabilities in high range resolution and robust anti-jamming, we propose and design a chipless UWB RFID tag for localization to resist the interference from the environment. In order to avoid interferences with these existing frequency bands, a compact coplanar waveguide (CPW)-fed ultra wideband (UWB)printed planar antenna with four band-notches is proposed and demonstrated. A bubble structure UWB antenna is designed first of all, based on which the parasitic element and the etching slot are added. Subsequently the design parameters of the resonator units that affect the resonator frequency based on both of the parasitic elements method and etching slots method are discussed. Simulation results show that both of the two methods cannot create the band-notch in the entire band of UWB, where the parasitic elements method cannot create a band-notch lower than 5GHz and the slot method can only create a band-notch with an upper limit of 8GHz. As a result, we discuss and design to combine both methods to achieve the any notchband in the entire UWB band. The low-profile antenna consists of a C-shape parasitic strip to generate a notched band at 8.01?8.55GHz for the ITU band,two C-shaped slots and an inverted U-shaped slot etched in the radiator patch to create three notched bands at 5.15?5.35GHz, 5.75?5.85GHz and 7.25?7.75GHz for filtering the WLAN and X-band satellite signals. Simulated and measured results both confirm that the proposed antenna has a broad bandwidth of 3.1 ? 12GHz with VSWR<2 and good omnidirectional radiation patterns with four notched-bands. Moreover, it is noted that the length of the parasitic/slot elements can control the band-notch frequency, which makes the multiple notchbans UWB antenna can be designed as a chipless tag. A chipless RFID tag can be designed by adding the number of the parasitic/slot elements and changing their length. For example, based on the designed planar bubble structure UWB antenna, we design a 5-bits chipless tag by adding different parasitic/slot elements, and simulation results show that the designed tag can be enconded.Thirdly, DV-hop algorithm is a well known localization algoritm in wsn, which has attracted many attentions in recent years. We analyse the factors that affect the localization accuracy of the DV-hop algorithm detailly, and propose two improved schemes for different sensor fields. As the DV-hop algorithm works by transforming the distance to all beacon nodes from hops to units of length measurement using the average size of a hop as a correction factor, the average hop distance and the hop-count between the destination and the source nodes are the two key factors that affect the localization accuracy. We investigate and optimize the parameters that affect these two factors, including the communication radius of the node, the number of beacon nodes and the number of the total nodes. For the hybrid optical and wireless sensor networks (O-WSNs) with a rectangular topology, we propose an improved DV-hop called Sub-Square Weighted DV-hop (SSW DV-hop) based on the optimization of the parameters. Both simulation and experiment results demonstrate that applying the SSW DV-hop algorithm in O-WSNs could significantly improve the localization accuracy, more than 65% improvement achieved. In the other hand, the DV-hop algorithm just uses the average distance of a hop and the hop counts between the beacon node and unknown node to calculate the distance,which doesn't make the best of the effect from the beacon nodes. As a result, we introduce the vector refinement feed-back method into the DV-hop localization which makes full use of the effect on the unknown node from different beacon nodes, called Vector Refinement DV-hop (VRDV-hop) algorithm. An action vector is assumed to evaluate the relationship between the unknown node and beacon nodes,indicating the difference between the hop-based distance and the Euclidean distance obtained by DV-hop. A weighted method based on the hops is introduced to eliminate the error in distance estimation accumulated with the increase of the path length from different beacon nodes. Parameters such as the number of the beacon node, the communication radius are discussed. Simulation and erperiment show that the proposed VRDV-hop algorithm outperforms than the DV-hop algorithm and the WDV-hop algoritm, 25% improvement achieved compared to their original basis.Finally, secure sensor localization in WSNs has received considerable attention with the rise of internet of things. For the presence of the malicious beacon nodes in WSN, we propose an efficient and secure range-free localization scheme called Outlier Elimination Vector Refinement DV-hop (OEVRDV-hop) algorithm. Based on the DV-hop algorithm, an outlier elimination method is introduced into the localization process by filtering inaccurate beacon nodes, where the HopSize of the beacon nodes are distinguished using the Pauta Criterion. A vector refinement feedback process is introduced to further improve the localization accuracy. Parameters that affect the localization accuracy (such as Value of the Pauta criterion, number of the malicious beacon nodes, bias of the malicious beacon nodes and number of the iteration, etc) are discussed. Simulation and experiment are carried out and compared to demonstrate the performance of the proposed OEVRDV-hop when the malicious beacon nodes are included, and also compared with the DV-hop and VRDV-hop algorithm works without/with malicious beacons. The localization accuracy of the DV-hop is not satisfactory due to the limited of the hardware and the effect from the environment,where the average localization error is 2.84m. What is more, the presence of the malicious beacon nodes deteriorates the localization accuracy, where the average localization error is up to 4.63m. The VRDV-hop has an outstanding accuracy whose average localization falls to 0.82m. However, it cannot work anymore in the presence of the malicious beacon nodes, the average localization error even increases to 4.34m. Fortunately, the proposed OEVRDV-hop is robust to overcome these problems and reduces the average localization error to 1.83m, an improvement of 55.2%, 153% and 137.2% compared to DV-hop, Attacked DV-hop and Attacked VRDV-hop, exhibits good performance on the attack-resistant. |
PDF Full Text Request