| Wireless sensor networks as a new information access and processing technology, can widely used in national defense and military, environmental monitoring, health care, space exploration and fighting terrorism and many other fields, be considered the one of the most important technology in twenty-first century. For most applications, the data which have no space-time identification has no use. Location has become essential to one of supporting technologies in sensor networks, which has been widespread concern of scholars home and abroad at present. Relatively perfect positioning technology such as GPS positioning technology, because by positioning time, positioning accuracy and the complex environmental conditions, can not be fully qualified for a full range of positioning required, particularly in the interior environment, it is almost entirely unavailable. Therefore, the study of high-precision positioning technology has drawn greater attention, the study is to use a lot of small, lightweight and low-cost sensors to form wireless sensor networks, and even the use of sensor fusion technology to achieve the purpose of high-precision positioning in the interior environment.The main job of this article has four parts:1. After introduce the important role of location technologies in applications of the wireless sensor network, have discussed some issue in the current study of wireless sensor networks, such as: lack of a more effective method of precise location, large-scale wireless sensor networks in the cumulative error and so on. Then cited four evaluation criteria about a Wireless sensor network self-positioning system and algorithm properties: positioning accuracy, scale, reference node density and node density. These evaluation criteria is a reference standard of the validity of a wireless sensor network system, they can inspect a sensor network whether or not be able to meet application requirements in a specific field.2. Then introduce a wireless sensor network architecture that is very typical. Behind the introduction of sensor nodes in the structure, focuses on the characteristics of wireless sensor networks and applications.3. This part is the theoretical basis of the full text. Focuses on three basic location algorithm: triangular measurement method, triangulation method, as well as maximum likelihood estimation method, the triangular measurement requires hardware support of ultrasound, because of the lack of such equipment, in this part of the back did not verify it. Introduce two kinds of mainstream location algorithm, range-based and the range-Free, because of implementation Part of this article are based on the distance of the RSSI (Received Signal Strength Indicator, received signal strength) location positioning algorithm , therefore focuses on RSSI technologies, including an empirical model and the three theoretical models, these models are described in signal strength and distance relationship. Range-Free algorithm require low hardware and without measuring the absolute inter-node distance or direction, at a rough estimate of position has a important role. Because of the higher complexity of triangular measurement and maximum likelihood estimation algorithm, which is very unfavorable to the sensor node whose computing resources are limited, In view of this, based on RSSI and centroid algorithm we raise three kinds of weighted centroid algorithms, reduced a lot of complexity. The same as trilateral and maximum likelihood estimation method, this algorithm is also dependent on the RSSI between blind node and the reference node to convert into distance, which is also a centralized location algorithm. Basic idea is that the RSSI value or distance between reference and blind node as the basis for calculating the weights of every reference node. By the weights value to explain the reference node how to influence centroid coordinate. Take advantage of weights to reflect reference node how to influence position of centroid, reflecting the intrinsic relationship between them. It is clear that the algorithm is very reasonable in theory, but RSSI value are easy to be strongly influenced by the environment, RSSI of the same distance is very different, RSSI measured is the same, in fact it may not be equal in distance, So we have introduced that make actual distance between reference nodes and their signal strength as a reference signal strength, make use of this information to amend the weights of reference nodes value, which were raised by Wang Yang in his doctoral thesis at University of Science and Technology of China in 2007. The weight of Reference node which only be considered the distance from reference nodes to the blind node ,and now use the actual distance and signal strength between reference nodes and the signal strength from reference nodes to the blind node as weights value of the reference nodes. Range-based and range-free algorithms have their the best scope of application, so these kinds of algorithm is summarized and concluded that discussed what kind of network environment can achieve better accuracy.4. This part focuses on RSSI-based location methods relying on the location the engine in CC2431 to calculate position of location node (also known as mobile nodes). The whole algorithm is based on the protocol stack ZigBee2006 platform. Discussed in detail the communication rules of three kinds of nodes (reference nodes, blind node, the gateway node) between them (9 kinds of string command). These string command defined the information that nodes in order to calculate the coordinates, For example, gateway nodes (also called the coordinator node) receive packets of coordinates location node sends, send data packet need for configuring reference node or blind nodes the data packet, Reference node to receive packet blind node request to collect RSSI average value, packet response to request of RSSI blind node sends. After blind nodes have collected all the parameters and coordinates, and save them into register MEASPARM and REFCOORD, which are used to calculate coordinates by location engine. Because of the algorithm of the location process were firmed in the location Engine, there are no way to know which algorithm are used for location, so we implemented the triangular measurement method and maximum likelihood estimation method to compare with the coordinates the location engine calculated, in the calculation of the algorithm the signal strength between nodes is important, but the sensor nodes only collect RSSI value, so we derived relationship between the physical distance, signal strength and RSSI value ,which aim at convert three kinds of values into each other. The results are that the coordinates error calculated by the triangular and the maximum likelihood estimation method are much larger than the location engine, with the increase in the number of reference node location error the maximum likelihood estimation method calculate decreases. On the ChengDu Wuxianlong platform, we implement the algorithm, the result is the same as what we expect, take the distance between the reference node and the signal strength as well as the signal strength from the reference node to blind node into account at the same time as the weights of the reference nodes calculate location error than the simple use distance between reference node and blind node t as weights of reference nodes, the error has been about 27% smaller. |
PDF Full Text Request