Genetic Algorithm Methods in Indoor Localization Based on Artificial Neural Networks
We proposed a novel fingerprinting method based on genetic algorithm (GA) to overcome the weakness of the artificial neural network (ANN) algorithm in searching global optimums and achieve precise localization.GA based on ANN (GA-ANN) can improve the accuracy and stability of conventional ANN algorithms by optimizing the weighs and biases of neural network.The localization accuracy has been improved by 158.52% and 19.96% respectively compared to the well-known back-propagation ANN (BP-ANN) and general regression neural network (GRNN).The fingerprinting method based on the GA-ANN method shows a huge advantage and potential in indoor localization technology.
genetic algorithm artificial neural network indoor localization fingerprinting method
Jichao Huang Liangxiao Tang Weixin Gai Xing Zhang
Institute of Microelectronics,Peking University, Beijing 100871, China
国内会议
成都
英文
1-5
2014-10-01（万方平台首次上网日期，不代表论文的发表时间）