WO2019051841A1

WO2019051841A1 - Method for determining filter coefficient and device therefor, and terminal

Info

Publication number: WO2019051841A1
Application number: PCT/CN2017/102108
Authority: WO
Inventors: 曾端; 王鑫山; 李国梁; 杨柯
Original assignee: 深圳市汇顶科技股份有限公司
Priority date: 2017-09-18
Filing date: 2017-09-18
Publication date: 2019-03-21
Also published as: CN107820677A; CN107820677B

Abstract

Provided are a method for determining a filter coefficient and device therefor, and a terminal. The method for determining a filter coefficient comprises: determining an estimated output of the filter according to an input signal of the filter at the current moment; determining an output error of the filter according to a target output and an estimated output of the filter; determining, according to the output error of the filter and a pre-determined first threshold, whether the current filter coefficient needs to be corrected; if it is determined that the current filter coefficient needs to be corrected, correcting the current filter coefficient to determine the filter coefficient at the next moment; and if it is determined that the current filter coefficient does not need to be corrected, directly using the current filter coefficient as the filter coefficient of the next moment, and thereby completely filtering out the interference or filtering out the interference to the greatest possible extent during filtering.

Description

Method for determining filter coefficient, device and terminal thereof

Technical field

The embodiments of the present application relate to the field of data processing technologies, and in particular, to a method, device, and terminal for determining filter coefficients.

Background technique

With the development of smart medical technology, ear channel thermometers, heart rate headphones and other terminal devices based on audio channel communication have been widely studied and applied. The traditional communication method based on the audio channel can achieve a low data transmission rate and cannot meet the application scenario of high-rate data exchange.

To this end, it is necessary to consider a high-rate, high-reliability communication method based on an audio channel. In order to achieve high-speed, high-reliability communication, how to ensure higher decoding success rate is especially important, and the decoding success rate is related to interference caused by channel noise and channel nonlinear factors, channel noise and channel nonlinear factors. The more thoroughly filtered, the higher the decoding success rate. Especially for different audio channels, the signal-to-noise ratio and channel nonlinearity of the signal will also change. For example, when playing music on the mobile phone with an audio application, selecting different sound effects will cause the audio channel to change, if the filter coefficient is usually If it is fixed, it is difficult to achieve thorough filtering of interference, which makes it difficult to ensure a high decoding success rate, and thus it is difficult to achieve high-speed and high-reliability communication, and cannot meet the application scenario of high-rate data exchange.

Summary of the invention

In view of this, one of the technical problems to be solved by the embodiments of the present invention is to provide a method for determining filter coefficients, a device thereof, and a terminal for overcoming or relieving the above-mentioned defects in the prior art.

An embodiment of the present application provides a method for determining a filter coefficient, including:

Determining an estimated output of the filter based on an input signal at a current moment of the filter;

Determining an output error of the filter based on a target output of the filter and an estimated output;

Determining whether the current filter coefficient needs to be corrected according to an output error of the filter and a predetermined first threshold;

If it is determined that the current filter coefficient needs to be corrected, the current filter coefficient is corrected to determine the filter coefficient at the next moment;

If it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next time.

Optionally, in an embodiment of the present application, modifying the current filter coefficient to determine a filter coefficient at a next moment includes:

Determining a correction amount of the current filter coefficient;

The current filter coefficient is modified according to the correction amount of the current filter coefficient to determine a filter coefficient at a next moment.

Optionally, in an embodiment of the present application, before determining the estimated output of the filter according to the input signal of the current time filter, the method further includes: according to a predetermined training sequence and the filter in the filter coefficient training phase The input signal determines the current filter coefficient used when filtering the input signal at the current moment of the filter.

Optionally, in an embodiment of the present application, determining, according to a predetermined training sequence and an input signal of the filter, a current filter coefficient used when performing filtering processing on an input signal of a current time of the filter includes:

Sampling a predetermined training sequence to obtain a target output of the filter during a filter coefficient training phase;

Determining an estimated output of the filter during a filter coefficient training phase based on an input signal of the filter in a filter coefficient training phase;

Determining an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase;

Determining whether training is required according to the output error of the filter in the filter coefficient training phase and the predetermined second threshold, and if so, training to obtain a current filter used for filtering the input signal of the current time of the filter coefficient.

Optionally, in an embodiment of the present application, the method further includes: if training is needed, determining whether the accumulated number of times of training exceeds a predetermined threshold of training times, and if not, jumping to a predetermined training sequence Sampling yields the target output of the filter at the filter coefficient training stage to continue training.

Optionally, in an embodiment of the present application, if the accumulated number of times of training exceeds a predetermined threshold of training times, the training ends.

Optionally, in an embodiment of the application, the training sequence is a random sequence or a pseudo random sequence.

Optionally, in an embodiment of the present application, the training sequence is selected from a data stream obtained by decoding an estimated output of the filter in a filtering stage.

Optionally, in an embodiment of the present application, whether the current filter coefficient needs to be corrected is indicated by a predetermined correction flag.

Optionally, in an embodiment of the present application, the predetermined training flag is used to indicate whether training is needed to obtain the current filter coefficient.

An embodiment of the present application provides an apparatus for determining a filter coefficient, including:

a filter, configured to determine an estimated output of the filter according to an input signal of a current moment of the filter;

An output error determining module, configured to determine an output error of the filter according to a target output of the filter and an estimated output;

a filter coefficient determining module, configured to determine, according to an output error of the filter and a predetermined first threshold, whether the current filter coefficient needs to be corrected, and if it is determined that the current filter coefficient needs to be corrected, the current filtering is performed The coefficient is modified to determine the filter coefficient at the next moment. If it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next moment.

Optionally, in an embodiment of the application, the filter coefficient determining module includes:

a revision amount determining unit, configured to determine a correction amount of the current filter coefficient;

And a correction unit, configured to modify the current filter coefficient according to the correction amount of the current filter coefficient to determine a filter coefficient at a next moment.

Optionally, in an embodiment of the present application, the method further includes: a training module, configured to determine, at a filter coefficient training stage, an input signal to a current moment of the filter according to a predetermined training sequence and an input signal of the filter The current filter coefficients used in the filtering process.

Optionally, in an embodiment of the application, the training module includes:

a target output determining unit, configured to sample a predetermined training sequence to obtain a target output of the filter in a filter coefficient training phase;

An estimated output determining unit configured to determine an estimated output of the filter during a filter coefficient training phase according to an input signal of the filter in a filter coefficient training phase;

An output error determining unit configured to determine an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase ;

a filter coefficient determining unit, configured to determine whether training is required according to an output error of the filter in the filter coefficient training phase and a predetermined second threshold, and if so, performing training to obtain an input signal of a current time of the filter The current filter coefficient used in the filtering process.

Optionally, in an embodiment of the present application, the method further includes: a training number determining module, configured to determine, when training is needed, whether the accumulated number of times of training exceeds a predetermined threshold of training times, and if not, the target The output determination unit continues to sample the predetermined training sequence to obtain a target output of the filter during the filter coefficient training phase to continue training.

An embodiment of the present application provides a terminal, including: a processor and a filter, where:

The filter is configured to determine an estimated output of the filter according to an input signal of the current moment;

The processor is configured to sample the received analog signal to obtain a sampling signal and use it as an input signal of the current moment of the filter;

The processor is further configured to determine an output error of the filter according to a target output of the filter and an estimated output, and determine whether a current filter is needed according to an output error of the filter and a predetermined first threshold. The coefficient is corrected;

The processor is further configured to: when the current filter coefficient needs to be corrected, correct the current filter coefficient to determine a filter coefficient at a next moment;

The processor is further configured to directly use the current filter coefficient as the filter coefficient of the next moment when it is determined that the current filter coefficient does not need to be corrected.

Optionally, in an embodiment of the present application, the filter is further configured to perform filtering processing on an input signal of the filter at a next moment according to a filter coefficient at a next moment.

Optionally, in an embodiment of the present application, the processor is further configured to perform synchronous frame header detection, Manchester decoding, and cyclic redundancy check processing on the estimated output of the filter.

In the technical solution of the embodiment of the present application, the estimated output of the filter is determined according to an input signal of a current moment of the filter; and the determined output is determined according to a target output of the filter and an estimated output. An output error of the filter; determining, according to the output error of the filter and the predetermined first threshold, whether the current filter coefficient needs to be corrected; if it is determined that the current filter coefficient needs to be corrected, the current filter is The coefficient is modified to determine the filter coefficient at the next moment; if it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next moment, so that the filtering process can be implemented according to The change of the data transmission channel adaptively corrects the current filter coefficients to completely filter out the interference or filter out the interference as much as possible, thereby ensuring a high decoding success rate, and finally achieving high-speed, high-reliability communication.

DRAWINGS

Some specific embodiments of the embodiments of the present application will be described in detail, by way of example, and not limitation, The same reference numbers in the drawings identify the same or similar parts. Those skilled in the art should understand that the drawings are not necessarily drawn to scale. In the figure:

1 is a schematic structural diagram of a communication system according to Embodiment 1 of the present application;

2 is a schematic flowchart of processing an original signal by a lower computer in Embodiment 2 of the present application;

FIG. 3 is a schematic diagram of the principle of encoding in the third embodiment of the present application,

4 is a schematic diagram showing the working flow of a second communication terminal as a host computer in Embodiment 4 of the present application;

5 is a schematic diagram of a filtering process in Embodiment 5 of the present application;

6 is a block diagram of filtering in Embodiment 6 of the present application;

7 is a schematic diagram of a training process in Embodiment 7 of the present application;

8 is a schematic diagram of determining a filter coefficient according to Embodiment 8 of the present application;

9 is a schematic structural diagram of an apparatus for determining a filter coefficient in Embodiment 9 of the present application;

FIG. 10 is a schematic structural diagram of a terminal in Embodiment 10 of the present application.

Detailed ways

Any technical solution for implementing the embodiments of the present invention necessarily does not necessarily need to achieve all of the above advantages at the same time.

For a better understanding of the technical solutions in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments are only a part of the embodiments of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art should be within the scope of protection of the embodiments of the present invention based on the embodiments in the embodiments of the present invention.

The specific implementation of the embodiments of the present invention is further described below in conjunction with the accompanying drawings.

In the technical solution of the following embodiments, the estimated output of the filter is determined according to an input signal of a current moment of the filter; and an output error of the filter is determined according to a target output of the filter and an estimated output. And determining, according to the output error of the filter and the predetermined first threshold, whether the current filter coefficient needs to be corrected; if it is determined that the current filter coefficient needs to be corrected, correcting the current filter coefficient to determine The filter coefficient of the next time; if it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next time. For example, when the audio channel changes, the signal-to-noise ratio and channel nonlinearity will change. The estimated output of the filter changes, further causing the output error of the filter to change, and then determining whether to correct the current filter coefficient according to the output error of the filter and the predetermined first threshold. Specifically, for example, when the statistical value of the output error of the filter exceeds the first threshold, it is determined that the current filter coefficient needs to be corrected to determine the filter coefficient at the next moment, and the filtering is performed according to the next moment in the filtering phase. The coefficients are filtered to achieve complete filtering of interference or to filter out interference as much as possible, so that the target output of the filter is as consistent as possible with the estimated output, ensuring a high decoding success rate, and finally achieving high-speed, high-reliability communication. .

In the following embodiments of the present application, the heart rate earphone is used as the first communication terminal, the smart phone is the second communication terminal, and the original signal is the original heart rate signal as an example for description. The first communication terminal and the second communication terminal transmit signals through the audio channel. The audio channel may specifically include an audio interface and an audio line.

In the above communication system including the first communication terminal and the second communication terminal, when the audio channel may change in different time periods, the channel characteristics may also change, such as the channel nonlinearity changes. For such a similar situation, in the process of high-speed data transmission rate communication between the first communication terminal and the second communication terminal, the filter coefficients are adaptively modified, thereby improving the demodulation success rate and achieving the required filtering requirements. Thorough filtering of interference.

However, it should be noted that, in the enlightenment of the following embodiments, the technical solution of the present application can be applied to any occasion requiring data processing, such as an ear thermometer and a smart terminal.

In order to facilitate understanding of the solution of the present application, in the following embodiments, a specific communication system to which the filter coefficient adjustment method of the present application is applied is first briefly described as an example. When applied in this specific communication system, filtering is only part of the process, and the analog signal is obtained after sampling, encoding, etc. before filtering, and after decoding, decoding is performed to obtain the decoded bit stream for use by the application. See the example description below.

Applying the filter coefficient update method of the present application to a specific communication system, the communication system includes: a first communication terminal, a second communication terminal, and a filter. As described above, the first communication terminal 101 is illustratively a heart rate. The earphone, in a specific application, the heart rate earphone as the lower computer, the second communication terminal 102 is illustratively an intelligent terminal, and the heart rate earphone cooperates with the intelligent terminal to achieve the purpose of heart rate detection, wherein the smart terminal is configured with a filter, see the figure 1. It is a schematic structural diagram of a communication system in Embodiment 1 of the present application.

In this embodiment, the first communication terminal is configured to acquire an original signal, process the original signal to obtain an analog signal, and transmit the analog signal to the second communication terminal; as described above, since the first communication terminal is a heart rate earphone Therefore, in this embodiment, the original signal is specifically a heart rate original signal, and the analog signal is specifically a heart rate analog signal.

Specifically, the first communication terminal (ie, the heart rate earphone) as the lower computer processes the heart rate original signal to obtain a heart rate analog signal, and an exemplary flow is shown in FIG. 2 , which is the lower machine pair original signal in the second embodiment of the present application. Schematic diagram of the processing flow; as shown in Figure 2, it specifically includes:

S201: Perform packet processing on the original signal.

In this embodiment, when the first communication terminal performs group packet processing on the original signal, each data packet includes a synchronization frame header, a data field, and a cyclic redundancy check bit.

In this embodiment, the synchronization frame header is used as a smart phone of the upper computer to detect a data packet. When the host computer When the sync frame header is detected, it is considered that a new data packet is detected before the subsequent decoding process can be started. The data fields may vary in different application environments, and may include data types, data lengths, packet serial numbers, and actual data.

S202. Encode the original signal processed by the group packet to obtain a digital signal.

In this embodiment, the first communication terminal uses the Manchester code to encode the original signal after the packet processing. Specifically, refer to FIG. 3, which is a schematic diagram of the principle of coding in the third embodiment of the present application. The original signal "0" after the packet processing is mapped to "10", and the original signal "1" is mapped to "01". The transmission time occupied by the corresponding bit of each original signal is T, and the bit time corresponding to each original signal after encoding is T/2, and the data transmission rate is 1/T bps.

S203. Perform modulation processing on the digital signal to obtain an analog signal.

In this embodiment, when the digital signal is modulated, the bit "0" corresponds to the first analog level, and the bit "1" corresponds to the second analog level. In a specific application scenario, for example, the first analog level is -V, and the second analog level is +V, and V represents the level of the level.

Processing by the second communication terminal samples the analog signal to obtain a sampling signal and transmits the sampling signal to a filter as an input signal of the filter; the filter is configured to determine the filtering according to an input signal of the filter The estimated output of the second communications terminal is further configured to determine an output error of the filter based on a target output of the filter and an estimated output, and further based on an output error of the filter and a predetermined a first threshold, determining whether correction of the current filter coefficient is needed; if yes, correcting the current filter coefficient to determine a filter coefficient at a next moment;

The filter is configured to filter the input signal of the filter at the next moment according to the filter coefficient at the next moment.

Specifically, in this embodiment, referring to FIG. 4, it is a schematic diagram of a working process of a second communication terminal as a host computer in Embodiment 4 of the present application; as shown in FIG. 4, it includes:

S401. Sample the received analog signal to obtain a sampling signal.

In this embodiment, the sampling parameters such as the sampling frequency and the data bit width are different depending on the application scenario.

S402. The filter performs filtering processing on the sampling signal.

In this embodiment, the filter is an FIR band-pass filter. If the current filter needs to be corrected during filtering, the host computer corrects the current filter coefficient to obtain the filter coefficient at the next moment, and the filter is based on the next filter. The filter coefficient at a moment filters the input signal at the next moment of the filter. The detailed filtering process can be seen in Figure 5 below.

S403. Perform synchronization frame header detection on the estimated output of the filter.

S404, it is determined whether the synchronization frame header is detected, and if so, step S405 is performed, and if not, then the process proceeds to step S401;

In this embodiment, the data of a certain length after the synchronization frame header is transmitted to the processing of step S405. A certain length should be sufficient to cover the entire packet, especially the data field. If the sync frame header is not detected, it jumps to step S401 to perform the next round of processing.

S405. Perform Manchester decoding on the estimated output of the filter after the synchronization frame header detection process.

Referring to the above-described encoding process, this step S405 performs the processing process opposite to the above encoding, from The decoded bit stream is obtained.

S406. Perform cyclic redundancy check processing on the decoded bit stream.

In this embodiment, the cyclic redundancy check process performs error detection on the decoded bit stream.

S407. Transmit the cyclic redundancy check result and the decoded bit stream to the application.

In this embodiment, the application program may refer to an application installed on the electronic terminal in conjunction with the ear thermometer and the heart rate earphone.

In this embodiment, the estimated output of the filter is the signal actually output by the filter after the filtering process.

The application determines the heart rate value calculation and the like based on the cyclic redundancy check result and the decoded bit stream.

Specifically, as shown in FIG. 5, it is a schematic diagram of a filtering process in Embodiment 5 of the present application; as shown in FIG. 6, is a block diagram of filtering in Embodiment 6 of the present application; as shown in FIG. 5 and FIG. 6, the following S501- S504 is a method for determining a filter coefficient or a correction algorithm. It should be noted that the current filter coefficient can be continuously trained according to a predetermined training sequence and an input signal of the filter. For details, refer to the following. Figure 7 shows.

It should be noted that, in the process of performing filtering, the current filter coefficient may be obtained by real-time training the initial filter coefficients according to a predetermined training sequence and an input signal of the filter, or may be obtained by using a training process before use. Existing current filter coefficients. For example, in a scenario, after the application restarts, the audio channel changes. If the filtering process is to be started, the initial filter coefficients need to be trained in real time to obtain the current filter coefficients; if only the application is paused, the audio is The channel does not change, and the initial filter coefficients need not be trained in real time to obtain the current filter, but only continue to use the current filter coefficients existing before the pause.

In this embodiment, the specific filtering process is as follows:

S501. Determine an estimated output of the filter according to a current filter coefficient and an input signal of the filter.

In this embodiment, it is assumed that the input signal x(n) at time n includes the target output d(n) of the filter and the interference noise v(n), as shown in the formula (1):

x(n)=d(n)+v(n) (1)

S502. Determine an output error of the filter according to a target output of the filter and an estimated output.

In this embodiment, if the filter uses an FIR filter, W _n represents the coefficient of the p-th order filter, and the formula (2) represents the filter coefficient of the p-th order filter:

W _n = [ω _n (0), ω _n (1), ..., ω _n (P)] ^T (2).

Estimated output of the filter stage filter

Generated by the convolution of the input signal X(n) of the filter and the filter coefficient W _n , as shown in equation (3):

In the above formula (3), X(n) = [x(n), x(n-1), ..., x _n (np)] ^T .

At time n, the output error e(n) of the filter is the difference between its target output and the estimated output, ie, as in equation (4):

As shown in equation (4), the magnitude of the output error e(n) reflects the proximity of the estimated output to the target output. The degree, for example, in a specific application scenario, if e(n) is smaller, it indicates that the estimated output is closer to the target output.

In the present embodiment, since the input signal X(n) of the filter is known, the target output d(n) can be obtained by pseudo-sampling the input signal x(n).

S503. Determine, according to the output error of the filter and the predetermined first threshold, whether the current filter coefficient needs to be corrected; if yes, execute step S504; if not, execute step S506: directly use the current filter coefficient as The filter coefficient of the next moment determines whether the communication between the upper computer and the lower computer ends. If not, the process goes to step S501 to continue processing the next input signal;

In this embodiment, when it is determined whether the current filter coefficient needs to be corrected, whether the statistical value of the output error of the filter exceeds a predetermined first threshold is determined, and if yes, the current filter coefficient needs to be performed. Corrected. The size of the first threshold may be adjusted according to the filtering accuracy requirement. If the estimated output is to be as consistent as possible with the target output, the first threshold may be smaller, otherwise the first threshold may be larger.

In this embodiment, a cost function ξ(n) may be defined according to the following formula (5). The value of the cost function is used as a statistical value of the output error. The smaller the statistical value, the closer the estimated output is to the target output.

In the above formula (5), where λ is a weighting factor for “forgetting” the old output error, the so-called “over old” can be understood as the earlier output error before the current filter coefficient, and the latest output is tracked in real time. The error, λ satisfies 0 < λ ≤ 1. The larger λ, the slower the response of the filter to the new output error, and the smaller the λ, the smaller the effect of the old output error.

In the above formula (5), the statistical value of the output error is obtained by performing a square sum of the output errors. In other embodiments, the sum of the absolute values of the output errors may also be statistically derived to obtain a statistical value of the output error.

In the above formula (5), the least squares method is taken as an example for description. In other embodiments, a least mean square (LMS) and a lattice algorithm may also be used.

S504. Determine a correction amount of the current filter coefficient, and correct the current filter coefficient according to the correction amount of the current filter coefficient to determine a filter coefficient at a next moment.

In this embodiment, the correction amount ΔW _{n of} the current filter coefficient may be specifically calculated according to the input signal X(n) and the output error e(n), and is calculated according to the input signal X(n) and the output error e(n). The technical processing of the correction amount ΔW _n is already in the prior art and will not be described here.

S505. The filter performs filtering processing on the input signal of the filter at the next moment according to the filter coefficient at the next moment.

In this embodiment, specifically, the current filter coefficient W _{n may} be corrected according to W _n+1 =W _n +ΔW _n to obtain the filter coefficient W _n+1 at the next moment, at the next moment of the filter. The input signal is used for filtering.

It should be noted that step S501 to step S504 in this embodiment can also be understood as a flow of adaptively correcting the filter coefficients.

In this embodiment, the filter coefficients of the next time are determined as a whole according to the above steps S503 and S504.

The process of training the initial filter coefficients to obtain the current filter coefficients can be as shown in FIG. As shown in FIG. 7 , it is a schematic diagram of a training process in the seventh embodiment of the present application. The initial filter coefficients are trained according to a predetermined training sequence and an input signal of the filter in the training phase to obtain a current filter coefficient. Specifically, it includes the following steps:

S701. Initialize a current filter coefficient to obtain an initial filter coefficient.

In this embodiment, for example, a specific value is directly given to the current filter coefficient to complete initialization, and an initial filter coefficient is obtained.

S702. The predetermined training sequence is sampled and the predetermined training sequence is sampled to obtain a target output of the filter in the training phase.

In this embodiment, the training sequence is a random sequence or a pseudo random sequence. The training sequence is selected from a data stream obtained by decoding an estimated output of the filter during a filtering phase. For example, the decoded bit stream obtained by the synchronization frame header detection and decoding is processed for the previous filtering loop.

In this embodiment, the target output D(n) for the training process is obtained by pseudo-sampling the modulated signal of the training sequence S(n). In this embodiment, in order to obtain the target output of the training phase filter, the sampling frequency used may be the same as the actual sampling frequency of the upper computer.

S703. The host computer samples the received analog signal to obtain a sampling signal and uses it as an input signal of the filter in the training phase;

As described above, in the present embodiment, the sampling frequency of the upper computer is the same as the sampling frequency used for the target signal of the filter obtained in the above step S702.

S704. Determine an estimated output of the filter in a training phase according to an initial filter coefficient of the filter and an input signal of the filter in a filter coefficient training phase.

The estimated output of the filter is obtained by referring to the following formula (6).

In formula (6), X(n)=[x(n), x(n-1),...,x _n (np)] ^T

S705. Determine an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase.

The output error E(n) of the filter described in the training phase is obtained by referring to the following formula (7):

S706. Determine, according to an output error of the filter in the filter coefficient training stage, and a predetermined second threshold, whether the initial filter coefficient needs to be trained, if yes, execute step S707; if not, execute the filter coefficient. Determine the method and reset the cumulative number of trainings;

In this embodiment, when it is determined whether the initial filter coefficient needs to be trained, whether the statistical value of the output error of the filter exceeds a predetermined second threshold according to the training phase is determined, and if yes, the initial filter is needed. The coefficients are trained.

The statistical values in this embodiment can be described with reference to the statistical values in FIG. 5 described above.

In this embodiment, the size of the second threshold may be determined according to historical empirical data, or may be flexibly adjusted according to the requirements of filtering accuracy. Specifically, the second threshold may be the same as or different from the first threshold. In a specific application, in order to obtain a more accurate current filter coefficient, the second threshold is specifically smaller than the first threshold.

S707. Determine a correction amount of an initial filter coefficient in the training phase;

In this embodiment, the correction amount ΔW _n ' of the initial filter coefficient may be specifically calculated according to the input signal x(n) of the filter and the output error E(n), for example, the input signal X(n) and the output. The error E(n) applies a least mean square (LMS) or a recursive least square (RLS) to calculate the correction amount ΔW _n '.

S708. Train the initial filter coefficients of the filter according to the correction amount of the initial filter coefficients to obtain current filter coefficients.

In this embodiment, specifically, the initial filter coefficient W _n ' can be corrected according to W′ _n+1 =W _n ′+ΔW _n ′ to obtain the modified filter coefficient W′ _n+1 , that is, the current filter. coefficients, i.e., the filter coefficient W after the correction _{'n + 1} as the filter of the current filter coefficient W _n in the embodiment.

It should be noted that, in the first training process, the initial filter coefficient W _n ' refers to the initializing filter coefficient obtained at the time of initial initialization, and after at least one training process, refers to the modified initializing filter coefficient.

S709. Determine whether the accumulated number of times of training exceeds a predetermined threshold of training times. If not, go to step S702 to continue to correct the initial filter coefficients; if the accumulated number of times of training exceeds a predetermined threshold of training times, Then the training of the initial filter coefficients is ended.

The training process shown in FIG. 7 above can also be understood as a process of continuously correcting the initial filter coefficients to obtain the current filter coefficients before the filter performs filtering.

On the basis of FIG. 7, a filter coefficient correction scheme is provided in combination with whether or not training processing is required, as shown in FIG. 8 , which is a schematic diagram of determining filter coefficients in the eighth embodiment of the present application;

S801, setting a training flag bit;

If the value of the training flag is set to 1, it indicates that the training process needs to be performed. If the value of the training flag is set to 0, the training process does not need to be performed;

For example, when the application restarts, the training flag is 1, and when the application is paused, the training flag is 0.

S802, determining, according to the training flag, whether the training process needs to be performed, for example, the value of the training flag is 1, and performing the step S803 after performing the training process of the above steps S701-S709; otherwise, directly performing step S803;

S803. The sampling signal obtained by sampling the analog signal is used as an input signal of the filter;

Referring to formula (3), the filtered signal is the estimated output of the filter.

S804, determining an output error of the filter and determining whether a statistical value of an output error of the filter exceeds a predetermined first threshold; if not, executing step S805; if yes, setting a correction flag flag to 1 And performing the above step S504, after which, go to step S805 again;

S805, determining whether the communication between the upper computer and the lower computer is over, and if so, then ending, otherwise proceeds to step 802;

In this embodiment, the value of the application flag bit can be specifically obtained and the communication between the upper computer and the lower computer is terminated according to the value of the application flag. For example, if the application flag bit is 1, the end is indicated.

In this embodiment, a training flag bit and a correction flag bit are set. The training flag bit is used to indicate whether the initial filter coefficient needs to be trained during the training process, and the correction flag bit is used to indicate whether it needs to be current in the filtering process. The filter coefficients are corrected.

It should be noted that, in other embodiments, the training flag bits may also be multiplexed to indicate whether the current filter coefficients need to be corrected during the filtering process, that is, the training flag bits are simultaneously multiplexed into the modified flag bits.

In another embodiment, the first communication terminal is a host computer, the second communication terminal is a lower computer, and the first communication terminal includes the filter, and the detailed communication process is similar to the embodiment shown in FIG. 1 above. .

It should be noted that, in the above embodiment, the filter may be a digital filter.

9 is a schematic structural diagram of an apparatus for determining a filter coefficient according to Embodiment 9 of the present application; as shown in FIG. 9, an estimated output of the filter is determined by a filter 901 according to an input signal of a current moment of a filter; and a filter coefficient is determined. The output error determination module 902 and the filter coefficient determination module 903 included in the apparatus are as follows:

An output error determining module 902, configured to determine an output error of the filter according to a target output of the filter and an estimated output;

The filter coefficient determining module 903 is configured to determine, according to the output error of the filter and the predetermined first threshold, whether the current filter coefficient needs to be corrected, and if it is determined that the current filter coefficient needs to be corrected, the current The filter coefficients are corrected to determine the filter coefficients at the next moment. If it is determined that the current filter coefficients need not be corrected, the current filter coefficients are directly used as the filter coefficients of the next moment.

Optionally, in any embodiment, the filter coefficient determining module 903 includes:

Optionally, in any embodiment, the method further includes: a training module, configured to determine, according to a predetermined training sequence and an input signal of the filter, a filter processing on an input signal of a current moment of the filter in a filter coefficient training phase The current filter coefficient to use.

Optionally, in any embodiment, the training module includes:

An estimated output determining unit configured to determine an estimated output of the filter during a filter coefficient training phase according to an input signal of the filter;

An output error determining unit, configured to determine an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase;

It should be noted that, in a specific application scenario, the estimated output determining unit may reuse the above estimation. The output output determining unit may multiplex the output error determining module, and the filter coefficient determining unit may multiplex the filter coefficient determining unit.

Optionally, in any embodiment, the method further includes: a training number determining module, configured to determine, when training is needed, whether the accumulated number of times of training exceeds a predetermined threshold of training times, and if not, jump to a predetermined The training sequence is sampled to obtain the target output of the filter during the filter coefficient training phase to continue training.

10 is a schematic structural diagram of a terminal in Embodiment 10 of the present application; as shown in FIG. 10, it includes: a filter 1001 and a processor 1002, where:

The filter 1001 is configured to determine an estimated output of the filter according to an input signal of a current moment of the filter;

The processor 1002 is configured to sample the received analog signal to obtain a sampling signal and use it as an input signal of the current moment of the filter;

The processor 1002 is further configured to determine an output error of the filter according to a target output and an estimated output of the filter, and determine, according to an output error of the filter and a predetermined first threshold, whether current filtering is needed. Correction of the coefficient of the device;

The processor 1002 is further configured to: when determining that the current filter coefficient needs to be corrected, correct the current filter coefficient to determine a filter coefficient at a next moment;

The processor 1002 is further configured to directly use the current filter coefficient as the filter coefficient of the next moment when it is determined that the current filter coefficient does not need to be corrected.

Optionally, in any embodiment, the filter 1001 is further configured to perform filtering processing on an input signal of the filter at a next moment according to a filter coefficient at a next moment.

Optionally, in any embodiment, the processor 1002 is further configured to perform synchronous frame header detection, Manchester decoding, and cyclic redundancy check processing on the estimated output of the filter.

The device embodiments described above are merely illustrative, wherein the modules described as separate components may or may not be physically separate, and the components displayed as modules may or may not be physical modules, ie may be located A place, or it can be distributed to multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. Those of ordinary skill in the art can understand and implement without deliberate labor.

Through the description of the above embodiments, those skilled in the art can clearly understand that the various embodiments can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware. Based on such understanding, portions of the above technical solutions that contribute substantially or to the prior art may be embodied in the form of a software product that may be stored in a computer readable storage medium, the computer readable record The medium includes any mechanism for storing or transmitting information in a form readable by a computer (eg, a computer). For example, a machine-readable medium includes read only memory (ROM), random access memory (RAM), magnetic disk storage media, optical storage media, flash storage media, electrical, optical, acoustic, or other forms of propagation signals (eg, carrier waves) , an infrared signal, a digital signal, etc., etc., the computer software product comprising instructions for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform the various embodiments or portions of the embodiments described Methods.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present application, and The present invention is described in detail with reference to the foregoing embodiments, and those skilled in the art should understand that the technical solutions described in the foregoing embodiments may be modified, or some of the technical features may be equivalent. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.

Those skilled in the art will appreciate that embodiments of the embodiments of the invention may be provided as a method, apparatus (device), or computer program product. Thus, embodiments of the invention may be in the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware. Moreover, embodiments of the invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

Embodiments of the invention are described with reference to flowchart illustrations and/or block diagrams of methods, apparatus, and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or FIG. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine for the execution of instructions for execution by a processor of a computer or other programmable data processing device. Means for implementing the functions specified in one or more of the flow or in a block or blocks of the flow chart.

The computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device. The apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.

These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device. The instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Claims

A method for determining a filter coefficient, comprising:

Determining an estimated output of the filter based on an input signal at a current moment of the filter;

Determining an output error of the filter based on a target output of the filter and an estimated output;

Determining whether the current filter coefficient needs to be corrected according to an output error of the filter and a predetermined first threshold;

If it is determined that the current filter coefficient needs to be corrected, the current filter coefficient is corrected to determine the filter coefficient at the next moment;

If it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next time.
The method of claim 1 wherein modifying the current filter coefficients to determine filter coefficients for the next moment comprises:

Determining a correction amount of the current filter coefficient;

The current filter coefficient is modified according to the correction amount of the current filter coefficient to determine a filter coefficient at a next moment.
The method according to claim 1, wherein determining the estimated output of the filter according to an input signal of the current time filter further comprises: according to a predetermined training sequence and the filter in a filter coefficient training phase The input signal determines the current filter coefficient used when filtering the input signal at the current moment of the filter.
The method according to claim 3, wherein determining the current filter coefficients used in filtering the input signal of the current time of the filter according to the predetermined training sequence and the input signal of the filter comprises:

Sampling a predetermined training sequence to obtain a target output of the filter during a filter coefficient training phase;

Determining an estimated output of the filter during a filter coefficient training phase based on an input signal of the filter in a filter coefficient training phase;

Determining an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase;

Determining whether training is required according to the output error of the filter in the filter coefficient training phase and the predetermined second threshold, and if so, training to obtain a current filter used for filtering the input signal of the current time of the filter coefficient.
The method according to claim 4, further comprising: if training is required, determining whether the accumulated number of times of training exceeds a predetermined threshold of training times, and if not, jumping to a predetermined training sequence Sampling yields the target output of the filter at the filter coefficient training stage to continue training.
The method according to claim 5, wherein the training is ended if the accumulated number of times of training exceeds a predetermined threshold of training times.
The method of claim 3 wherein said training sequence is a random sequence or Pseudo-random sequence.
The method of claim 3 wherein said training sequence is selected from a data stream obtained by decoding an estimated output of said filter during a filtering phase.
The method of claim 1 wherein the predetermined correction flag is used to indicate whether correction of the current filter coefficients is required.
The method according to claim 3, wherein the current filter coefficients are obtained by indicating whether training is required by a predetermined training flag.
A device for determining a filter coefficient, comprising:

An output error determining module, configured to determine an output error of the filter according to a target output of the filter and an estimated output;

a filter coefficient determining module, configured to determine, according to an output error of the filter and a predetermined first threshold, whether the current filter coefficient needs to be corrected, and if it is determined that the current filter coefficient needs to be corrected, the current filtering is performed The coefficient is modified to determine the filter coefficient at the next moment. If it is determined that the current filter coefficient does not need to be corrected, the current filter coefficient is directly used as the filter coefficient of the next moment;

Wherein the estimated output of the filter is determined by the filter based on the input signal at the current moment of the filter.
The apparatus according to claim 11, wherein the filter coefficient determining module comprises:

a revision amount determining unit, configured to determine a correction amount of the current filter coefficient;

And a correction unit, configured to modify the current filter coefficient according to the correction amount of the current filter coefficient to determine a filter coefficient at a next moment.
The apparatus according to claim 11, further comprising: a training module, configured to determine an input signal to a current time of the filter according to a predetermined training sequence and an input signal of the filter in a filter coefficient training phase The current filter coefficients used in the filtering process.
The device of claim 13, wherein the training module comprises:

a target output determining unit, configured to sample a predetermined training sequence to obtain a target output of the filter in a filter coefficient training phase;

An estimated output determining unit, configured to determine an estimated output of the filter in a filter coefficient training phase according to an input signal of the filter in a filter coefficient training phase;

An output error determining unit configured to determine an output error of the filter in a filter coefficient training phase according to a target output of the filter in a filter coefficient training phase and an estimated output of the filter in a filter coefficient training phase ;

a filter coefficient determining unit, configured to determine whether training is required according to an output error of the filter in the filter coefficient training phase and a predetermined second threshold, and if so, performing training to obtain an input signal of a current time of the filter The current filter coefficient used in the filtering process.
The apparatus according to claim 14, further comprising: a training number determining module, configured to determine whether the accumulated number of times of training exceeds a predetermined number of training times when training is required, and if not, the target The output determination unit continues to sample the predetermined training sequence to obtain a target output of the filter during the filter coefficient training phase to continue training.
A terminal, comprising: a processor and a filter, wherein:

The filter is configured to determine an estimated output of the filter according to an input signal of the current moment;

The processor is configured to sample the received analog signal to obtain a sampling signal and use it as an input signal of the current moment of the filter;

The processor is further configured to determine an output error of the filter according to a target output of the filter and an estimated output, and determine whether a current filter is needed according to an output error of the filter and a predetermined first threshold. The coefficient is corrected;

The processor is further configured to: when the current filter coefficient needs to be corrected, correct the current filter coefficient to determine a filter coefficient at a next moment;

The processor is further configured to directly use the current filter coefficient as the filter coefficient of the next moment when it is determined that the current filter coefficient does not need to be corrected.
The terminal according to claim 16, wherein the filter is further configured to filter the input signal of the filter at the next moment according to the filter coefficient at the next moment.
The terminal according to claim 16, wherein the processor is further configured to perform synchronous frame header detection, Manchester decoding, and cyclic redundancy check processing on the estimated output of the filter.