[time]
2022-08-24T11:05:53+02:00

[track]
7

[team_name]
WHU_CIRpos

[team_institution]
LIESMARS, WHU

[logolink]


[team_members]
Jingfeng Mao, Jinbin Liu, Xiaodong Gong, Delong Liu, Fei Yin, Xuanfan Lv

[reference_person]
Jingfeng Mao

[reference_email]
15827289218@163.COM

[description_short]
We propose an error regression method based on support vector machine and convolutional neural network, and on this basis, we use weighted least squares and adaptive UKF filtering to obtain the final localization results. In the support vector machine method, we perform autonomous support vector machine regression with parameter selection optimization by using PSO-SVR regression method. We also use an adaptive algorithm that allows the regression filter to autonomously select the corresponding time-domain features. (For the regression features, we mainly consider Energy of the received signal, Maximum amplitude of the received signal, Rise time, Mean excess delay, RMS delay spread, Kurtosis, etc.). After the model regression of the ranging error is achieved, the ranging value is corrected and a weighting model is set for the error regression value, and the weighted least squares and adaptive Kalman filter localization solution is performed, and CNN method is also tried .



[description_long_link]
https://pan.baidu.com/s/12r0XM8iK5sZ4Qiy4FrILEw?pwd=ya7f 提取码: ya7f 

[publish_check_]
true

[results_check_]
true

[data_check_]
true

[pdf_check_]
true

[video_check_]
true