Research On Indoor Positioning Technology Based On ZigBee Sensor Network

Now, ZigBee wireless technology is a strong impetus to the development ofwireless personal area networks, as a new wireless network communicationtechnology of low-cost, low-power consumption, and large-capacity. ZigBeetechnology is based on IEEE802.15.4standard, and a standard used in wirelessdetection and control of the global wireless communications, now it has been widelyused in consumer electronics, health care, agriculture and intelligent remote controland other fields. At current, relatively mature outdoor positioning technology hasbeen in the presence defects of poor signal quality, high-end cost, and high powerconsumption in indoor environment. So using the features about low network cost,low power, and with a significant security encryption of ZigBee to solute indoorpositioning, is the hot spot of positioning question.Paper describes the development of indoor positioning technology of domesticand aboard; and gives the content and objectives. We study and analyze the ZigBeewireless sensor network technology, and select TI’s CC2530chip to design andachieve a complete Zigbee wireless positioning system, and study positioningtechnology of moving target under complex indoor environment, and do experimentalverification and test analysis on this positioning system platform.We use fingerprint database positioning algorithm to solver target locationcombine with indoor-dimensional positioning requirements for accuracy. As thetraditional fingerprint algorithm has disadvantages about heavy workload of theprocess in building a database, and RF data is susceptible to noise interference, weproposed a method of generating virtual fingerprint database partitions based onpolynomial interpolation, and use particle filter technique in processing estimatedpre-processing results at positioning stage, and introduce the multilayer neuralnetwork into filtering algorithms to solve the problem of irregular distribution ofRSSI effectively. Optimized algorithm can quickly generate simple fine-grainedpositioning information database, and its positioning error is less than1.14m.According to the three-dimensional characteristics of indoor positioning research, we use the weighted-nearest neighbor algorithm and Bayesian positioning algorithmsto get target location based on path loss model. For uncertainty electromagnetic ofsignals ranging theoretical model parameters in traditional algorithms, we combinewith the actual measurement environment, and proposed a parameter estimationmethod for electromagnetic signal path loss model based on evolutionary strategy.Experiments show that in a large enough three-dimensional space, the signal strengthindicate deviation of estimation model is less than3.7dBm within20m, comparedwith the measured values, and average positioning error can be controlled in3.4m orless by optimized algorithm.