The Location Algorithm Based on Square-root Cubature Kalman Filter
Liu Ying
1, a
, Su Junfeng
1, b
, Zhu Mingqiang
1, c
1
School of Electronics and Information Engineering, Beijing Jiao tong University, Beijing, China
a
liuying@bjtu.edu.cn
b
sjf19851214@163.com
c
mqzhu@bjtu.edu.cn
Keywords: indoor localization, Received Signal Strength Indication (RSSI), Square-root Cubature
Kalman filter (SCKF), localization error.
Abstract. When wireless signal is used for indoor localization, due to the complex environment of
multi-path effects, there is no consistent relationship between the signal strength received by the
receiving nodes and the distance from the receiving nodes to the receiving nodes, so there is a larger
localization error for the Received Signal Strength Indication (RSSI) in the indoor environment. A
new received signal strength indicator parameter estimation algorithm based on square-root cubature
kalman filter is proposed in this paper, this algorithm converts the RSSI localization problem into
the parameter vector estimation problem of nonlinear equations, which utilizes the Square-root
Cubature Kalman filter (SCKF) to estimate the target’s position and the RSSI channel
attenuation parameter simultaneously, and uses dynamic channel parameter to correct the node’s
position in real – time. The experimental results demonstrate that the RSSI parameter estimation
algorithm based on SCKF is more accurate than the traditional method based on the least-square curve
fitting in the indoor wireless localization.
Introduction
With the development of wireless communication networks, the localization of network nodes is
becoming increasingly important. In the application of network load balancing, network topology and
the route based on the location information, location information of the nodes are very useful and
important for the realization of the communication and collaboration process [1]. Localization method
based on Received Signal Strength Indicator (RSSI) is widely used in mobile wireless networks. In the
existing wireless network, mobile nodes location function can be achieved by adding the signal strength
detection unit and appropriate location algorithm. There are quite different for RSSI methods between
the indoor and outdoor. Because of walls, floors and other obstacles in the indoor, the reflection,
refraction and diffraction will lead to large random fluctuations for the RSSI measurements. The key
problem for RSSI localization is how to accurately estimate the radio propagation model parameters,
in order to improve localization accuracy.
Considering the affection of a variety of factors for RSSI measurements in the indoor environment,
a large number of beacon nodes (BN) can be used as reference nodes for localization, and then the
channel attenuation parameter were estimated. This method requires a large number of beacon nodes,
which means a high cost [2]. If only fewer beacon nodes were used, the RSSI distribution data in each
beacon must be pre-stored. At this case it is needed to build a huge database of RSSI distribution data.
Once the indoor environment changed such as indoor facilities location, the database of RSSI
distribution date should be rebuild [3, 4]. We found that there is a transmission distance threshold in
indoor environment by detecting experiment. When the transmission distance is less than this threshold,
there is a consistent relationship between the RSSI measurement and the transmission distance. At this
case the method based on least-square curve fitting is suitable for RSSI localization [5]. The location
algorithm based on least-square curve fitting coverts channel attenuation parameter estimation and