CN-122028059-A - Communication base station investigation system, method, equipment and storage medium

CN122028059ACN 122028059 ACN122028059 ACN 122028059ACN-122028059-A

Abstract

The invention discloses a communication base station investigation system, a method, a storage medium and electronic equipment, wherein the method comprises a multi-source sensing unit, a data processing unit, an intelligent decision unit and a history address selection unit, wherein the multi-source sensing unit is used for acquiring basic data of each candidate address at different sampling moments, the data processing unit is used for inputting the basic data of each candidate address at different moments, inputting a pre-trained deep learning model to obtain attribute labels and prediction basic data of each candidate address, transmitting the prediction basic data of an initial candidate address with the attribute labels being first labels to the intelligent decision module, and inputting the prediction basic data of the initial candidate address and the history address selection data as inputs to the pre-trained address selection evaluation model to screen the optimal address of the communication base station from the initial candidate address.

Inventors

LIU SHUANG
CHENG HONGBO
LIU FENG
LI ZHUO

Assignees

中国移动通信集团设计院有限公司
中国移动通信集团有限公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1.A communication base station survey system, comprising: The multi-source sensing unit is used for acquiring basic data of each candidate address at different sampling moments, wherein the basic data at least comprises positioning data, environment parameter data and network quality data; The data processing unit is used for inputting basic data of each candidate address at different moments, inputting a pre-trained deep learning model to obtain attribute labels and prediction basic data of each candidate address, and transmitting the prediction basic data of which the attribute labels are initial candidate addresses of first labels to the intelligent decision module, wherein the attribute labels comprise first labels and second labels, the first labels are used for indicating that the corresponding candidate addresses can be used as initial candidate addresses of the communication base station, and the first labels are used for indicating that the corresponding candidate addresses cannot be used as initial candidate addresses of the communication base station; the intelligent decision unit is used for taking the prediction basic data and the historical address selection data of the initial candidate address as inputs, and inputting a pre-trained address selection evaluation model so as to screen the most preferred address of the communication base station from the initial candidate address.
2. The system of claim 1, wherein the deep learning model comprises a convolutional neural network CNN and a recurrent neural network RNN; The CNN comprises three convolution layers and a maximum pooling layer which are sequentially connected, wherein the three convolution layers are connected through a ReLU activation function, and the last convolution layer in the three convolution layers is connected with the maximum pooling layer; The RNN comprises a bidirectional long-short-term memory network LSTM layer, a Softmax layer and a full-connection layer, wherein the bidirectional LSTM layer is used as an input layer of the RNN to be connected with the maximum pooling layer, the Softmax layer and the full-connection layer are used as output layers of the RNN to be connected with the bidirectional LSTM layer, the Softmax layer is used for outputting attribute labels of candidate addresses, and the full-connection layer is used for outputting prediction basic data of the candidate addresses.
3. The system of claim 1, wherein the multi-source sensing unit comprises a positioning module, a multi-modal sensor module, and an interference monitoring module; the positioning module is used for collecting positioning data of the corresponding candidate address; The multi-mode sensor module is used for collecting environmental parameter data of the corresponding candidate address; the interference monitoring module is used for collecting network quality data of the corresponding candidate addresses.
4. The system of claim 3, wherein the multi-modal sensor module comprises a temperature sensor, a humidity sensor, and an illumination sensor; the temperature sensor is used for collecting temperature data in the environmental parameter data of the corresponding candidate address; The humidity sensor is used for collecting humidity data in the environment parameter data of the corresponding candidate address; the illumination sensor is used for collecting illumination data in the environment parameter data of the corresponding candidate address.
5. The system according to any one of claims 1 to 4, wherein the intelligent decision unit comprises a storage module and a decision module; The storage module is used for storing the predicted basic data of the initial candidate address and the history address selection data, wherein the history address selection data comprises the address of at least one history communication base station and the basic data of the corresponding history communication base station; the decision module is used for taking the prediction basic data and the historical address selection data of the initial candidate address as input and inputting a pre-trained address selection evaluation model so as to screen the most preferred address of the communication base station from the initial candidate address.
6. The system of claim 5, wherein the site selection evaluation model is constructed based on a fusion genetic algorithm and a particle swarm optimization algorithm.
7. The system according to claim 1, characterized in that the data processing unit is specifically configured to: Carrying out standardization processing on the basic data of each candidate address at different moments to obtain the standardized basic data of each candidate address at different moments, wherein the standardization processing comprises data cleaning processing and normalization processing; Taking standardized basic data of each candidate address at different moments as input, and inputting a pre-trained deep learning model to obtain attribute labels and prediction basic data of each candidate address; and transmitting the prediction basic data of which the attribute label is the initial candidate address of the first label to an intelligent decision module.
8. A method of communication base station investigation, comprising: Basic data of each candidate address at different sampling moments is obtained, wherein the basic data at least comprises positioning data, environment parameter data and network quality data; The method comprises the steps of taking basic data of each candidate address at different moments as input, inputting a pre-trained deep learning model to obtain attribute labels and prediction basic data of each candidate address, and transmitting the prediction basic data of an initial candidate address with the attribute labels being first labels to an intelligent decision module, wherein the attribute labels comprise first labels and second labels, the first labels are used for indicating that the corresponding candidate address can be used as the initial candidate address of a communication base station, and the first labels are used for indicating that the corresponding candidate address cannot be used as the initial candidate address of the communication base station; and taking the prediction basic data and the historical address selection data of the initial candidate address as inputs, and inputting a pre-trained address selection evaluation model to screen the optimal address of the communication base station from the initial candidate address.
9. A computer-readable storage medium, having stored thereon a computer program or instructions which, when run on a computer, cause the method of claim 8 to be performed.
10. An electronic device, comprising: And a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of claim 8.

Description

Communication base station investigation system, method, equipment and storage medium Technical Field The present application relates to the field of 6G, and in particular, to a communication base station investigation system, method, apparatus, and storage medium. Background The communication base station investigation is a key link before base station construction, and various factors such as topography, traffic conditions, power supply, surrounding environment and the like are required to be investigated and evaluated, so that network layout optimization, signal quality improvement and construction cost control are directly affected. In the prior art, the communication base station investigation mainly depends on an on-site investigation method, a geographic information system technology, a remote sensing technology, an unmanned aerial vehicle aerial photography technology and the like, and related patents relate to functions of Beidou data transmission, connection switching optimization, interference monitoring and the like, but have obvious defects in the aspects of technology integration and base station investigation site selection. The prior art fails to effectively integrate a 6G communication technology, a Beidou satellite navigation system and an Internet of things technology, so that the base station investigation positioning accuracy is limited, the data acquisition dimension is single, the data processing delay is higher, the intelligent decision making capability of base station investigation and site selection is lacked, and the requirement of high-quality construction of a communication network cannot be met. Therefore, there is a need for a communication base station survey system with high accuracy positioning, real-time data processing, and intelligent decision making capabilities. Disclosure of Invention The application provides a communication base station investigation system, a method, equipment and a storage medium, which are used for solving the problems of limited positioning precision data, single acquisition dimension, higher data processing delay, insufficient intelligent decision-making capability of base station investigation and site selection and the like in the prior art. In a first aspect, a communications base station survey system is provided, comprising: The multi-source sensing unit is used for acquiring basic data of each candidate address at different sampling moments, wherein the basic data at least comprises positioning data, environment parameter data and network quality data; The data processing unit is used for inputting basic data of each candidate address at different moments, inputting a pre-trained deep learning model to obtain attribute labels and prediction basic data of each candidate address, and transmitting the prediction basic data of which the attribute labels are initial candidate addresses of first labels to the intelligent decision module, wherein the attribute labels comprise first labels and second labels, the first labels are used for indicating that the corresponding candidate addresses can be used as initial candidate addresses of the communication base station, and the first labels are used for indicating that the corresponding candidate addresses cannot be used as initial candidate addresses of the communication base station; the intelligent decision unit is used for taking the prediction basic data and the historical address selection data of the initial candidate address as inputs, and inputting a pre-trained address selection evaluation model so as to screen the most preferred address of the communication base station from the initial candidate address. Optionally, the deep learning model includes a convolutional neural network CNN and a cyclic neural network RNN; The CNN comprises three convolution layers and a maximum pooling layer which are sequentially connected, wherein the three convolution layers are connected through a ReLU activation function, and the last convolution layer in the three convolution layers is connected with the maximum pooling layer; The RNN comprises a bidirectional long-short-term memory network LSTM layer, a Softmax layer and a full-connection layer, wherein the bidirectional LSTM layer is used as an input layer of the RNN to be connected with the maximum pooling layer, the Softmax layer and the full-connection layer are used as output layers of the RNN to be connected with the bidirectional LSTM layer, the Softmax layer is used for outputting attribute labels of candidate addresses, and the full-connection layer is used for outputting prediction basic data of the candidate addresses. Optionally, the multi-source sensing unit comprises a positioning module, a multi-mode sensor module and an interference monitoring module, wherein the positioning module is used for acquiring positioning data of the corresponding candidate address, the multi-mode sensor module is used for acquiring environment parameter data of