A Study Of Indoor Positioning Algorithms Based On RSSI

In recent years, the indoor positioning technology has become one of the most popular studied objects since it has been implemented in many fields thanking to the rapid development of Wireless Sensor Network (WSN). The key of the indoor positioning technology is the positioning algorithm. A well designed algorithm can smooth the channel noise; get a higher accuracy, while using less network resources. According to the faults of two original algorithms, this paper proposes methods to optimize them respectively.With five reference nodes at least, can linearization algorithm get an acceptable accuracy. With Taylor Series Expansion, we can overcome this shortcoming. First, we give the blind node an initial coordinate, then we expand the group of binary quadratic based on RSSI with Taylor Series at the point of initial coordinate, remove quadratic and higher, at last,we apply iteration algorithm to estimate the real coordinate of blind node. Compared to the original, this new method can get a very good accuracy with only three reference nodes.We present an optimization to smooth the divergence of the center algorithm which occurs when it has been interfered by a grave noise. The optimization exploits all intersection points got from solving group of binary quadratic, including the real and imaginary, to estimate the blind node's coordinate. Simulation concludes that it can enhance the robust of the algorithm with no more reference nodes. Further more, to remove the fault of the least squares estimation (LSE), we introduce weighted least squares estimation(WLSE) to enhance the accuracy.Finally, according to different conditions, this paper studies several algorithms to solve the velocity of the blind node which always is in the WSN, and applies the Kalman filter to estimate the locus and velocity of the blind node.