Limits...
Artificial neural network for location estimation in wireless communication systems.

Chen CS - Sensors (Basel) (2012)

Bottom Line: In a wireless communication system, wireless location is the technique used to estimate the location of a mobile station (MS).Artificial neural networks (ANN) are widely used techniques in various areas to overcome the problem of exclusive and nonlinear relationships.The numerical analysis and simulation results show that the proposed algorithms can obtain more precise location estimation under different NLOS environments.

View Article: PubMed Central - PubMed

Affiliation: Department of Information Management, Tainan University of Technology, Yongkang District, Tainan, Taiwan. t00243@mail.tut.edu.tw

ABSTRACT
In a wireless communication system, wireless location is the technique used to estimate the location of a mobile station (MS). To enhance the accuracy of MS location prediction, we propose a novel algorithm that utilizes time of arrival (TOA) measurements and the angle of arrival (AOA) information to locate MS when three base stations (BSs) are available. Artificial neural networks (ANN) are widely used techniques in various areas to overcome the problem of exclusive and nonlinear relationships. When the MS is heard by only three BSs, the proposed algorithm utilizes the intersections of three TOA circles (and the AOA line), based on various neural networks, to estimate the MS location in non-line-of-sight (NLOS) environments. Simulations were conducted to evaluate the performance of the algorithm for different NLOS error distributions. The numerical analysis and simulation results show that the proposed algorithms can obtain more precise location estimation under different NLOS environments.

No MeSH data available.


Average location error versus the upper bound of NLOS errors.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3376586&req=5

f6-sensors-12-02798: Average location error versus the upper bound of NLOS errors.

Mentions: The number of hidden neurons is determined through experimentation. If there are too few hidden neurons, it will cause a bigger error. Increasing the number of hidden neurons can alleviate this situation, but it will also affect the speeds of convergence simultaneously, and the computing would be almost no help in reducing NLOS errors after exceeding a certain number of neurons. The general rules for choosing the number of neurons in the hidden layer are: (i) 0.5(·p + q), (ii) p, (iii) 2·p + 1, (iv) 3·p + 1, where p and q are the input and output hidden neurons, respectively [35]. Figure 5 shows the RMS error obtained with different number of hidden layer neurons. One can see the RMS error converged to the same minimum value for various hidden layer neurons. The main factor of affecting the accuracy of MS location is not the numbers of hidden-layer neurons. Because of the satisfactory prediction performance, the number of hidden neurons is set to 0.5 · (p + q). In order to avoid increasing the computation load, we use the proposed algorithm with 0.5 · (p + q) hidden neurons and 1,000 epochs for both training and testing data in the following simulations. From Figures 4 and 5, we can find out that the positioning precision of the SCG, Rprop, and LM algorithm is better than CGF and CGP algorithm, especially in harsh NLOS environments. Based on the ability of estimating the neural network structure stated above, we apply the SCG, Rprop and LM algorithms to predict MS location after training period. Figure 6 shows the effect of various methods used with upper bound of NLOS error on the average location error. It is clear that as the upper bound of NLOS error increases, the average location error increases. Because of the square range-differencing operations involved, LLOP can mitigate the NLOS error. In comparison with LLOP’s reasonably results, TSA leads to less accurate results. The proposed algorithm is significantly more effective in radiolocation accuracy than TSA, LLOP and RSA, especially in severe NLOS conditions. It can be observed that the proposed algorithm can reduce the RMS errors effectively and estimate the MS location accurately.


Artificial neural network for location estimation in wireless communication systems.

Chen CS - Sensors (Basel) (2012)

Average location error versus the upper bound of NLOS errors.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3376586&req=5

f6-sensors-12-02798: Average location error versus the upper bound of NLOS errors.
Mentions: The number of hidden neurons is determined through experimentation. If there are too few hidden neurons, it will cause a bigger error. Increasing the number of hidden neurons can alleviate this situation, but it will also affect the speeds of convergence simultaneously, and the computing would be almost no help in reducing NLOS errors after exceeding a certain number of neurons. The general rules for choosing the number of neurons in the hidden layer are: (i) 0.5(·p + q), (ii) p, (iii) 2·p + 1, (iv) 3·p + 1, where p and q are the input and output hidden neurons, respectively [35]. Figure 5 shows the RMS error obtained with different number of hidden layer neurons. One can see the RMS error converged to the same minimum value for various hidden layer neurons. The main factor of affecting the accuracy of MS location is not the numbers of hidden-layer neurons. Because of the satisfactory prediction performance, the number of hidden neurons is set to 0.5 · (p + q). In order to avoid increasing the computation load, we use the proposed algorithm with 0.5 · (p + q) hidden neurons and 1,000 epochs for both training and testing data in the following simulations. From Figures 4 and 5, we can find out that the positioning precision of the SCG, Rprop, and LM algorithm is better than CGF and CGP algorithm, especially in harsh NLOS environments. Based on the ability of estimating the neural network structure stated above, we apply the SCG, Rprop and LM algorithms to predict MS location after training period. Figure 6 shows the effect of various methods used with upper bound of NLOS error on the average location error. It is clear that as the upper bound of NLOS error increases, the average location error increases. Because of the square range-differencing operations involved, LLOP can mitigate the NLOS error. In comparison with LLOP’s reasonably results, TSA leads to less accurate results. The proposed algorithm is significantly more effective in radiolocation accuracy than TSA, LLOP and RSA, especially in severe NLOS conditions. It can be observed that the proposed algorithm can reduce the RMS errors effectively and estimate the MS location accurately.

Bottom Line: In a wireless communication system, wireless location is the technique used to estimate the location of a mobile station (MS).Artificial neural networks (ANN) are widely used techniques in various areas to overcome the problem of exclusive and nonlinear relationships.The numerical analysis and simulation results show that the proposed algorithms can obtain more precise location estimation under different NLOS environments.

View Article: PubMed Central - PubMed

Affiliation: Department of Information Management, Tainan University of Technology, Yongkang District, Tainan, Taiwan. t00243@mail.tut.edu.tw

ABSTRACT
In a wireless communication system, wireless location is the technique used to estimate the location of a mobile station (MS). To enhance the accuracy of MS location prediction, we propose a novel algorithm that utilizes time of arrival (TOA) measurements and the angle of arrival (AOA) information to locate MS when three base stations (BSs) are available. Artificial neural networks (ANN) are widely used techniques in various areas to overcome the problem of exclusive and nonlinear relationships. When the MS is heard by only three BSs, the proposed algorithm utilizes the intersections of three TOA circles (and the AOA line), based on various neural networks, to estimate the MS location in non-line-of-sight (NLOS) environments. Simulations were conducted to evaluate the performance of the algorithm for different NLOS error distributions. The numerical analysis and simulation results show that the proposed algorithms can obtain more precise location estimation under different NLOS environments.

No MeSH data available.