Macrocell Path-Loss Prediction Using Artificial Neural Networks

被引:150
作者
Ostlin, Erik [1 ]
Zepernick, Hans-Jurgen [1 ]
Suzuki, Hajime [2 ]
机构
[1] Blekinge Inst Technol, S-37225 Ronneby, Sweden
[2] Commonwealth Sci Ind Res Org, Informat & Commun Technol Ctr, Marsfield, NSW 2122, Australia
关键词
Artificial neural network (ANN); backpropagation; feed-forward network; field strength measurements; Okumura-Hata (OH) model; path-loss prediction; point to area; Recommendation ITU-R P.1546; PROPAGATION;
D O I
10.1109/TVT.2010.2050502
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
This paper presents and evaluates artificial neural network (ANN) models used for macrocell path-loss prediction. Measurement data obtained by utilizing the IS-95 pilot signal from a commercial code-division multiple-access (CDMA) mobile network in rural Australia are used to train and evaluate the models. A simple neuron model and feed-forward networks with different numbers of hidden layers and neurons are evaluated regarding their training time, prediction accuracy, and generalization properties. Furthermore, different backpropagation training algorithms, such as gradient descent and Levenberg-Marquardt, are evaluated. The artificial neural network inputs are chosen to be distance to base station, parameters easily obtained from terrain path profiles, land usage, and vegetation type and density near the receiving antenna. The path-loss prediction results obtained by using the ANN models are evaluated against different versions of the semi-terrain based propagation model Recommendation ITU-R P. 1546 and the Okumura-Hata model. The statistical analysis shows that a non-complex ANN model performs very well compared with traditional propagation models with regard to prediction accuracy, complexity, and prediction time. The average ANN prediction results were 1) maximum error: 22 dB; 2) mean error: 0 dB; and 3) standard deviation: 7 dB. A multilayered feed-forward network trained using the standard backpropagation algorithm was compared with a neuron model trained using the Levenberg-Marquardt algorithm. It was found that the training time decreases from 150 000 to 10 iterations, while the prediction accuracy is maintained.
引用
收藏
页码:2735 / 2747
页数:13
相关论文
共 35 条
[1]  
ADAWI NS, 1988, IEEE T VEH TECHNOL, V37, P3
[2]  
[Anonymous], P15462 ITUR
[3]  
[Anonymous], P15463 ITUR
[4]  
Demuth H., 2003, NEURAL NETWORK TOOLB
[5]  
Foresee F. D., 1997, P INT C NEUR NETW IC, V3, P1930, DOI DOI 10.1109/ICNN.1997.614194
[6]  
Hagan M. T., 1997, Neural network design
[7]  
Hagan M.T., 1999, Proceedings of the 1999 American control conference, V3, P1642, DOI DOI 10.1109/ACC.1999.786109
[8]   TRAINING FEEDFORWARD NETWORKS WITH THE MARQUARDT ALGORITHM [J].
HAGAN, MT ;
MENHAJ, MB .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1994, 5 (06) :989-993
[9]  
Haykin S. S., 1994, Neural Networks: A Comprehensive Foundation
[10]  
Horn R.A., 2012, Matrix analysis, DOI [10.1017/CBO9780511810817, DOI 10.1017/CBO9780511810817]