阅读说明：本技术 用于预测修复的通信信道的比特率的方法和装置 (Method and apparatus for predicting bit rate of a repaired communication channel ) 是由 N·杜普伊斯 P·迪里克斯 于 2020-01-22 设计创作，主要内容包括：实施例涉及一种用于预测修复的通信信道的比特率的方法和装置。该方法可以包括：-生成数据集,数据集针对多个通信信道指定：-受损害影响的通信信道的信道频率响应；以及-未受所述损害影响的通信信道的信道频率响应；-基于数据集,训练机器学习模型,机器学习模型被配置用于基于受损的通信信道的信道频率响应来预测修复的通信信道的信道频率响应。(Embodiments relate to a method and apparatus for predicting a bit rate of a repaired communication channel. The method can comprise the following steps: -generating a data set, the data set specifying for a plurality of communication channels: -a channel frequency response of the communication channel affected by the impairment; -a channel frequency response of a communication channel not affected by said impairment; -training a machine learning model based on the data set, the machine learning model being configured for predicting a channel frequency response of the repaired communication channel based on the channel frequency response of the impaired communication channel.)

1. An apparatus comprising at least one processor and at least one memory, the at least one memory storing instructions, the at least one memory and the instructions configured to, with the at least one processor, cause the apparatus to perform:

-generating a data set specifying, for a plurality of communication channels:

-a channel frequency response of the communication channel affected by the impairment; and

-a channel frequency response of the communication channel unaffected by the impairment, specified by a parameter representing the relation between the channel attenuation on a logarithmic scale and the square root of the frequency;

-training a machine learning model based on the data set, the machine learning model configured for predicting a channel frequency response of a repaired communication channel based on a channel frequency response of a compromised communication channel, wherein training the machine learning model comprises determining an error, the error representing:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-a comparison between the predicted value of the parameter and the allowed value of the parameter.

2. The apparatus of claim 1, wherein generating the data set comprises: at least one of a channel frequency response of a communication channel affected by the impairment and a channel frequency response of a corresponding communication channel not affected by the impairment is determined based on circuit simulations.

3. The apparatus of claim 1, wherein training the machine learning model comprises determining an error that represents:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-an intersection between the predicted channel frequency response and the expected channel frequency response.

4. The apparatus of claim 1, wherein the machine learning model comprises a convolutional neural network comprising at least two convolutional layers.

5. The apparatus of claim 1, wherein the memory and instructions are further configured to cause the apparatus to: the data set is generated based on modem measurement effects typically observed in a measurement profile.

6. The apparatus of claim 1, wherein the memory and instructions are further configured to cause the apparatus to: a channel frequency response of a corresponding repaired communication channel is determined based on a channel frequency response measured on the impaired communication channel using a trained machine learning model.

7. The apparatus of claim 6, wherein the memory and instructions are further configured to cause the apparatus to: determining a bit rate or bit rate improvement for the repaired communication channel based on the channel frequency response of the corresponding repaired communication channel.

8. The apparatus of claim 1, wherein the instructions and memory are further configured to cause the apparatus to: the trained machine learning model is deployed in another device.

9. A computer-implemented method, comprising:

-generating a data set specifying, for a plurality of communication channels:

-a channel frequency response of the communication channel affected by the impairment; and

-a channel frequency response of the communication channel unaffected by the impairment, specified by a parameter representing the relation between the channel attenuation on a logarithmic scale and the square root of the frequency;

-training a machine learning model based on the data set, the machine learning model configured for predicting a channel frequency response of a repaired communication channel based on a channel frequency response of a compromised communication channel, wherein training the machine learning model comprises determining an error, the error representing:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-a comparison between the predicted value of the parameter and the allowed value of the parameter.

10. The method of claim 9, comprising: the trained machine learning model is deployed in another device.

11. The method of claim 9, wherein generating the data set comprises: at least one of a channel frequency response of a communication channel affected by the impairment and a channel frequency response of a corresponding communication channel not affected by the impairment is determined based on circuit simulations.

12. The method of claim 9, wherein training the machine learning model comprises determining an error that represents:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-an intersection between the predicted channel frequency response and the expected channel frequency response.

13. The method of claim 9, comprising: the data set is generated based on modem measurement effects typically observed in a measurement profile.

14. A computer program comprising instructions for performing the method according to claim 9 when the instructions are executed by a computer.

15. An apparatus obtained by the method of claim 10, comprising at least one processor and at least one memory storing instructions configured to, with the at least one processor, cause the apparatus to perform:

a channel frequency response of a corresponding repaired communication channel is determined based on a channel frequency response measured on the impaired communication channel using a trained machine learning model.

Technical Field

Embodiments of the invention relate to the field of telecommunications. In particular, embodiments of the present invention relate to a method and apparatus for predicting the bit rate of a repaired communication channel.

Background

In multi-tone communication systems such as DS L, it may be used to fill in the appropriate number of bits on different frequency carriers.

When a communication channel is affected by impairments, the channel frequency response and the achievable bit rate may be affected. Repairing the communication channel by removing impairments may improve the channel frequency response and achievable bit rate. However, the improved bit rate of the repaired communication channel may vary from one channel to another, and therefore the cost/benefit ratio of the corresponding repair actions may differ. Accordingly, it is desirable to predict the channel frequency response and/or achievable bit rate of a repaired communication channel for a compromised communication channel for which repair may be planned. Known prediction methods rely on, for example, curve fitting techniques, interpolation ….

Disclosure of Invention

It is therefore an object of embodiments of the present invention to propose a method and apparatus for predicting the channel frequency response of a communication channel which do not show the inherent drawbacks of the prior art.

Accordingly, embodiments relate to an apparatus comprising means configured for:

-generating a data set, the data set specifying for a plurality of communication channels:

-a channel frequency response of the communication channel affected by the impairment; and

-a channel frequency response of a communication channel not affected by said impairment;

-training a machine learning model based on the data set, the machine learning model being configured for predicting a channel frequency response of the repaired communication channel based on the channel frequency response of the impaired communication channel.

In some embodiments, generating the data set comprises: at least one of a channel frequency response of a communication channel affected by the impairment and a channel frequency response of a corresponding communication channel not affected by the impairment is determined based on circuit simulations.

In some embodiments, in the data set, the channel frequency response of the communication channel unaffected by the impairment is specified based on a parameter representing the relationship between the channel attenuation on a logarithmic scale and the square root of the frequency.

In some embodiments, training the machine learning model includes determining an error, the error representing:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-a comparison between the predicted value of the parameter and the allowed value of the parameter.

In some embodiments, training the machine learning model includes determining an error, the error representing:

-a comparison between the predicted channel frequency response and the expected channel frequency response; and

-an intersection between the predicted channel frequency response and the expected channel frequency response.

In some embodiments, the machine learning model includes a convolutional neural network including at least two convolutional layers.

In some embodiments, the apparatus is further configured to generate the data set based on modem measurement effects typically observed in the measurement profile.

In some embodiments, the apparatus is further configured to determine a channel frequency response of the corresponding repaired communication channel based on a channel frequency response measured on the impaired communication channel using the trained machine learning model.

In some embodiments, the apparatus is further configured to determine a bit rate or bit rate improvement for the repaired communication channel based on a channel frequency response of the corresponding repaired communication channel.

In some embodiments, the apparatus is further configured to deploy the trained machine learning model in another apparatus.

In some embodiments, the apparatus further comprises at least one processor and at least one memory, the at least one memory storing instructions, the at least one memory and the instructions configured to, with the at least one processor, cause the apparatus at least to perform, at least in part, the functions discussed above.

Drawings

The above and other objects and features of the invention will become more apparent and the invention itself will be best understood by reference to the following description of an embodiment taken in conjunction with the accompanying drawings, in which:

FIGS. 1 and 2 are graphs of channel frequency response;

FIG. 3 is a block diagram of a communication network;

fig. 4 is a flow chart of a method performed in the communication network of fig. 3;

FIG. 5 is a block diagram of a machine learning model configured for predicting a channel frequency response of a repaired communication channel based on a channel frequency response of an impaired communication channel;

fig. 6 is a flow chart of a method performed in the communication network of fig. 3;

fig. 7 is a structural view of a device used in the network of fig. 3.

Embodiments are also directed to a computer-implemented method, comprising:

-generating a data set, the data set specifying for a plurality of communication channels:

-a channel frequency response of the communication channel affected by the impairment; and

-a channel frequency response of a communication channel not affected by said impairment;

-training a machine learning model based on the data set, the machine learning model being configured for predicting a channel frequency response of the repaired communication channel based on the channel frequency response of the impaired communication channel.

In some embodiments, the method comprises: the trained machine learning model is deployed in another device.