Method for estimating interference in global navigation satellite system receiver

Abstract

A method for estimating interference in a Global Navigation Satellite System (GNSS) receiver is provided. The method is performed in a receiver and comprises: receiving a GNSS signal at a radio frequency band and processing the received signal at an intermediate frequency band; collecting, at the receiver, a set of parameters based on the received GNSS signal; and obtaining a likelihood value using a machine learning model trained for the receiver, where the set of parameters is an input to the machine learning model, the likelihood value is an output of the machine learning model, and the likelihood value is a number between 0 and 1 and corresponds to a likelihood that the receiver is disturbed in the intermediate frequency band.

Description

Technical Field

[0001] The present disclosure relates to a method for estimating jamming in a global navigation satellite system receiver and a method for training a machine learning model to estimate jamming. The present invention also relates to a device comprising a global navigation satellite system receiving unit and a processing unit. Background Art

[0002] The global navigation satellite system (GNSS) has been widely used to obtain positioning and timing information, which can be further used for various applications such as navigation, tracking, remote control, etc. In the GNSS system, multiple GNSS satellites above the earth's atmosphere continuously broadcast their GNSS signals at radio frequencies. Devices equipped with GNSS receiving functions can receive GNSS signals from GNSS satellites for further processing.

[0003] Accurate positioning information based on GNSS signals is generally expected to be used in various applications. However, since GNSS signals are generally attenuated when they reach the GNSS receiver, interference now occurs more and more frequently and commonly, and may seriously disrupt the GNSS signals. Interference may occur intentionally or unintentionally, and may be caused by different types of interfering signals (such as continuous wave signals, narrowband signals, broadband signals, etc.). Interference may also occur at different frequencies within the range of GNSS transmission frequencies. During the operation of the GNSS receiver, the GNSS receiver may encounter interference with any type of interfering signal at any unknown time. In the presence of interference, the reliability of positioning information based on GNSS signals cannot be guaranteed.

[0004] Therefore, what is desirable but challenging is a timely and lightweight estimate of interference so that the GNSS receiver can take timely countermeasures when interference occurs. This challenge is not well addressed in conventional solutions. Summary of the invention

[0005] An object to be achieved is to provide an improved concept for estimating interference in a GNSS receiver in a timely and lightweight manner.

[0006] According to the present disclosure, a GNSS receiver is capable of receiving GNSS signals at different radio frequency (RF) bands, down-converting and further processing the GNSS signals at corresponding intermediate frequency (IF) bands. The receiver may be interfered by any type of interference signal (e.g., transmitted by an illegal source). In this case, if any interference signal falls within the range of the RF band used to receive the GNSS signal, it may interfere with the received GNSS signal.

[0007] The improved concept is based on the idea that the receiver collects a set of parameters reflecting information of the received GNSS signals. In particular, the parameters are measured specifically for the RF band and the corresponding IF band of those GNSS signals received and processed at the RF and IF bands. The band-specific parameters are input to a machine learning model, for example, a stateless and lightweight logistic regression model, and the machine learning model has been trained specifically for the receiver. The machine learning model outputs a probability value indicating the probability of the receiver being interfered at a specific RF and IF band.

[0008] If the likelihood is deemed high, the receiver can perform countermeasures to eliminate the effects of the interference, for example, by performing interference mitigation techniques, by ceasing to receive GNSS signals from the affected RF band, or by ceasing to trust position, velocity, and time, PVT, solutions generated based on the interfered GNSS signals, etc.

[0009] By means of the frequency band specific parameters, the interference estimation can indicate whether the RF and IF frequency bands used for the received GNSS signals are affected by interference, so that the receiver can decide whether to continue using the GNSS signals at the affected frequency bands.

[0010] Since the machine learning model is stateless, interference can be estimated in a timely manner at the receiver, so that the effects of interference can be eliminated without delay. Interference estimation can be performed at any time during receiver operation, even at receiver startup. Interference estimation can quickly reflect changes in any existing interference, and quickly detect newly emerging interference.

[0011] Since the machine learning model is lightweight, receivers with constraints and computational resources (e.g., embedded GNSS receivers in mobile devices or IoT devices) are able to perform interference estimation locally at the receiver.

[0012] Furthermore, considering that different types of receivers may have different hardware characteristics as well as firmware and software properties, each machine learning model is trained specifically for one type of receiver, thereby optimizing each machine learning model for a specific type of receiver to achieve more accurate interference estimation.

[0013] According to the present disclosure, a method for estimating interference in a GNSS receiver includes: the receiver receives a GNSS signal at an RF band and processes the received signal at an IF band corresponding to the RF band, wherein the GNSS signal is transmitted on a carrier frequency at the RF band and the GNSS signal is down-converted to the IF band; the receiver collects a set of parameters at the receiver based on the received GNSS signal; and the receiver uses a machine learning model trained for the receiver to obtain a likelihood value, wherein the set of parameters is an input to the machine learning model, the likelihood value is an output of the machine learning model, and the likelihood value is a number between 0 and 1 and corresponds to the likelihood of the receiver being interfered with in the intermediate frequency band.

[0014] In various implementations of the method, the machine learning model is a logistic regression model. The logistic regression model is stateless and lightweight for timely estimation of interference.

[0015] In an example implementation of the method, the set of parameters includes the skewness of the RF spectrum at the RF band; the gain of the RF amplifier in the receiver; the skewness of the IF spectrum at the IF band; the variance of the IF spectrum at the IF band; the variance at the IF band under the IF histogram; the average value of the collapsed modulus-1kHz spectrum at the IF band, wherein the collapsed modulus-1kHz spectrum is selected from frequency components with modulus-1kHz values ​​in the spectrum around the IF; and the number of GNSS signals whose range tracking and / or velocity tracking is aborted by the receiver due to high correlation with noise. The parameters are measured for the RF and IF bands used by the receiver, so that the interference estimate can reflect the interference status of the RF and IF bands used. It has been found that these selected parameters constitute a reasonable and effective method according to the improved concept for estimating interference in a GNSS receiver. However, further parameters may be included.

[0016] In some implementations of the method, the receiver applies interference mitigation to the received GNSS signals before the receiver collects the set of parameters. In this case, the interference estimation is performed after the interference mitigation and may indicate whether the interference mitigation effectively eliminates the interference.

[0017] In some implementations, the method further includes: the receiver determines a threshold value, wherein the threshold value is a number between 0 and 1; the receiver determines the interference state as an interfered state when the likelihood value is greater than the threshold value, and determines the interference state as a non-interfered state when the likelihood value is less than or equal to the threshold value. By using the threshold value, the receiver can quickly determine the interference state and react to the interference state.

[0018] The present disclosure also provides a training method for training a machine learning model according to an improved concept of estimating interference in a GNSS receiver in a timely and lightweight manner. The trained model can be used in a method for estimating interference. The training is implemented by a computing device for estimating interference in a GNSS receiver. The training method includes: the receiver receives GNSS signals at multiple RF bands and processes the received signals at multiple corresponding IF bands, wherein the received GNSS signals include interfered GNSS signals and uninterrupted GNSS signals; the receiver records a set of parameters corresponding to the interfered GNSS signals and the uninterrupted GNSS signals at the corresponding IF bands; and the computing device trains the machine learning model based on the recorded set of parameters and the corresponding GNSS signals at the corresponding IF bands. With the aid of the training method, a single machine learning model is obtained for estimating interference at all IF bands, and is optimized for the receiver and receivers of the same receiver type.

[0019] In some implementations of the training method, recording the set of parameters includes: the receiver recording the set of parameters reflecting information that the received GNSS signals within the corresponding IF frequency band are interfered with or not interfered with.

[0020] In an example implementation of the training method, the set of parameters includes the parameters described above in the method for estimating interference in a GNSS receiver. In the training method, when the receiver receives GNSS signals at multiple RF frequency bands and processes the received signals at multiple corresponding IF frequency bands, the parameters are thus corresponding measurements at multiple RF frequency bands and multiple IF frequency bands.

[0021] In some implementations of the training method, training the machine learning model includes: a computing device setting an interference state of the GNSS signal at a corresponding intermediate frequency, wherein the interference state is a state in which the GNSS signal is interfered with or a state in which the GNSS signal is not interfered with; and the computing device training the machine learning model based on the recorded parameters and the corresponding interference state of the GNSS signal at the corresponding intermediate frequency.

[0022] In some implementations, the training method also includes a computing device simulating an interference signal and applying the simulated interference signal to at least a portion of the GNSS signal to be received by the receiver, wherein the simulated interference signal is a different type of signal including a continuous wave signal, a narrowband signal, and a wideband signal, and the simulated interference signal has different power levels, different frequencies, and different durations.

[0023] The present disclosure also provides a device according to an improved concept for estimating interference in a GNSS receiver in a timely and lightweight manner. The device includes a GNSS receiving unit and a processing unit. The device is configured to receive a GNSS signal at an RF band and process the received signal at an IF band; collect a set of parameters at the device; and use a machine learning model trained for the device to obtain a likelihood value, wherein the set of parameters is an input to the machine learning model, the likelihood value is an output of the machine learning model, and the likelihood value is a number between 0 and 1 and corresponds to the likelihood of the device being interfered in the IF band.

[0024] In various implementations of the device, the machine learning model is a logistic regression model.

[0025] In some implementations of the device, the set of parameters includes the parameters described above in the method for estimating interference in a GNSS receiver.

[0026] In some implementations, the device is further configured to apply interference mitigation to the received GNSS signals prior to collecting the set of parameters.

[0027] In some implementations, the device is further configured to: receive additional GNSS signals at multiple RF frequency bands and process the additional signals at multiple IF frequency bands, wherein the additional GNSS signals include interfered GNSS signals and undisturbed GNSS signals; and record another set of parameters to be used as a training data set, wherein the another set of parameters corresponds to the interfered and undisturbed additional GNSS signals at the corresponding IF frequency bands.

[0028] From the various implementations described above in conjunction with the method and the training method for estimating interference in a GNSS receiver, further implementations and developments of the apparatus will become apparent to those skilled in the art.

[0029] According to one embodiment of the improved concept, a computer program product comprises instructions that can be stored in a preferably non-transitory computer-readable storage medium, which instructions enable a computer system with one or more processors to perform a method according to one of the above-described implementations.

[0030] Furthermore, the computer system may have one or more processors and a storage medium in which computer program instructions are stored, such that the one or more processors can execute the method according to one of the above-described embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The improved concept will be explained in more detail below with the aid of the accompanying drawings. Throughout the accompanying drawings, elements and functional blocks having the same or similar functions have the same reference numerals. Therefore, their description does not need to be repeated in the following drawings.

[0032] In the figure:

[0033] Figure 1 A flow chart showing a method for estimating interference in a GNSS receiver;

[0034] Figure 2 An example of a logistic regression model is shown;

[0035] Figure 3 Examples of likelihood values ​​are shown;

[0036] Figure 4 A flowchart showing a method for training a machine learning model; and

[0037] Figure 5 An example implementation of a device is shown.

[0038] List of reference numbers

[0039] 100, 200 Method

[0040] Steps 101-103, 201-203

[0041] x, x1, x2, x3 input variables

[0042] w1, w2, w3 weights

[0043] y output variable

[0044] RECV receiving unit

[0045] PROC Processing Unit DETAILED DESCRIPTION

[0046] Figure 1 A flow chart illustrating an example implementation of a method 100 for estimating interference in a GNSS receiver is shown.

[0047] As an example, the receiver may be any kind of electronic device and / or electronic module with GNSS reception and processing capabilities. For example, the receiver may be a terminal device, a computer device, a module embedded in a sensor device, etc. The receiver may have constrained power and computing resources.

[0048] The GNSS may be any kind of constellation positioning system with which the receiver is capable of operating, such as GPS, GLONASS, BeiDou, Galileo, and the like.

[0049] In step 101 , a receiver receives a GNSS signal at an RF frequency band and processes the received signal at an IF frequency band.

[0050] For example, GNSS satellites continuously broadcast GNSS signals. When one or more GNSS satellites are visible to the receiver, the receiver can receive GNSS signals from the visible GNSS satellites. Carrier frequencies and corresponding RF bands are designated for sending different types of GNSS signals. For example, GPS L1C / A signals use a carrier frequency of 1575.42 MHz, Beidou B1I signals use a carrier frequency of 1561.098 MHz, and so on. Thus, the receiver will receive different types of GNSS signals at different RF bands centered on their carrier frequencies.

[0051] It is obvious that Figure 1 The example implementation of method 100 in shows an example case where the receiver receives one type of GNSS signal using the same carrier frequency. In the case where the receiver receives multiple types of GNSS signals using corresponding carrier frequencies, method 100 may be performed for each type of GNSS signal using the corresponding carrier frequency.

[0052] Referring back to step 101, after the receiver receives the GNSS signal at an RF frequency band centered on the corresponding carrier frequency, the GNSS signal is further processed in the processing chain. For example, the GNSS signal is received by an antenna or antenna array in the receiver and can be amplified by an RF amplifier (e.g., an RF programmable gain amplifier PGA). The amplified GNSS signal at the RF frequency band can be shifted to the corresponding IF frequency band by a down converter. The IF GNSS signal can be converted into a digital signal for further signal processing steps, such as acquisition, tracking, navigation processing, etc.

[0053] Obviously, in step 101 and the following steps of method 100, the carrier frequency, the RF frequency band, and the IF frequency band correspond to each other, and they belong to the same GNSS signal in different processing steps. For ease of reading, the RF frequency band and the IF frequency band belonging to the same GNSS signal are respectively referred to as in-use RF frequency band and in-use IF frequency band, and are collectively referred to as in-use frequency bands.

[0054] When there is interference, the received GNSS signal may be interfered. For example, when there is an interference signal at the RF frequency band in use, the received signal may therefore contain the original GNSS signal transmitted from the GNSS satellite and the interference signal. Along the processing chain, the interference signal may also be contained in the IF signal at the IF frequency band in use. In this case, the receiver is interfered at the IF frequency band in use. Subsequent signal processing results may be unreliable.

[0055] Interference signals can be various types of unknown signals, such as continuous wave, CW, signals, narrowband signals, wideband signals, etc. All types of interference signals can hinder the acquisition and tracking of GNSS signals to produce reliable navigation results. Interference signals can have various transmission power levels, various frequencies or durations.

[0056] During the reception and processing chain, certain interference mitigation techniques may be applied. For example, antennas may utilize spatial characteristics to block certain interfering signals, and notch filters may attenuate unwanted narrowband interfering signals. The following steps of method 100, i.e., steps 102 and 103, may be performed after any interference mitigation techniques are performed. In this case, the interference estimate may indicate whether the interference mitigation effectively eliminates the interference, and whether there is any residual interference after the interference mitigation.

[0057] In step 102, the receiver collects a set of parameters at the receiver based on the received GNSS signals. As an example implementation, the set of parameters reflects measurements and information specific to the frequency band in use of the received GNSS signals.

[0058] For example, the group of parameters may include the measurement result of the skewness of the RF spectrum at the RF frequency band in use, the measurement result of the gain of the RF amplifier used in the receiver, the measurement result of the skewness of the IF spectrum at the IF frequency band in use, the measurement result of the variance of the IF spectrum at the IF frequency band in use, and the measurement result of the variance of the IF histogram at the IF frequency band in use. The corresponding calculation of the skewness and variance of the spectrum and the variance of the frequency histogram is known to those skilled in the art, so it is not described in detail here.

[0059] The set of parameters may also include the average value of the collapsed modulus-1kHz spectrum at the IF band in use. To calculate this parameter, the frequency components of the signal at the IF band in use are analyzed. Frequency components with the same modulus-1kHz value are organized into subsets, for example, a subset of all frequency components starting from 50Hz with the same modulus-1kHz value will contain frequency components of 50Hz, 1050Hz, 2050Hz, etc., and another subset of all frequency components starting from 100Hz with the same modulus-1kHz value will contain frequency components of 100Hz, 1100Hz, 2100Hz, etc. For each subset, find the frequency component with the maximum energy. Then, select the found frequency component with the maximum energy in each subset to establish the collapsed spectrum. This parameter is the average value of the collapsed spectrum.

[0060] The parameters of the average value of the collapsed spectrum reflect in particular information about possible CW interfering signals. If any mitigation technique targeting mitigation of CW interfering signals is applied, for example by time domain filtering, the parameters provide information about any residual CW interfering signals.

[0061] The set of parameters may also include a number of GNSS signals whose tracking was aborted by the receiver due to high correlation with noise. For example, for GNSS ranging, the receiver performs distance tracking and / or velocity tracking based on the received GNSS signals. In the recent past, for example, during the interval in which the parameters were collected, the number of GNSS signals whose distance tracking and / or velocity tracking was aborted due to high correlation with noise is collected. The parameters may also reflect information about possible CW interfering signals.

[0062] Several of the parameters listed above are generated with respect to the frequency band in use of the received GNSS signal. In this way, the set of parameters collected is specific to the frequency band in use. As an example implementation of method 100, additional parameters may be collected at the receiver and used with the parameters listed above.

[0063] In step 103, the receiver inputs the set of parameters into the machine learning model and obtains a likelihood value output from the machine learning model. The likelihood value is a number between 0 and 1 and corresponds to the likelihood that the receiver is interfered at the IF band in use.

[0064] As an example implementation, the machine learning model is a logistic regression model. A logistic regression model typically takes input variables and gives the probability of a binary outcome. Figure 2 An example of a logistic regression model is shown. In the example model, x1, x2, and x3 are the three inputs to the model; w1, w2, and w3 are the weights of the model, each weight corresponding to a corresponding input; and y is the output of the model. When the model is used, the weights have been trained so that the output y can be calculated from the inputs and weights.

[0065] Referring back to step 103, the input to the machine learning model is the parameters listed above in step 102. For example, when using Figure 2 When the logistic regression model shown is used as a machine learning model, since seven parameters are listed above in step 102, these seven parameters can be x1 to x7 to be input to the logistic regression model. The logistic regression model has been trained so that each input parameter has a training weight for calculating the output, i.e., one of w1 to w7. The output of the machine learning model (e.g., Figure 2 y) in is a likelihood value, or in other words, a probability number between 0 and 1, which indicates the probability of being one of the two interference states.

[0066] Figure 3An example of a likelihood value that can be output from a machine learning model is shown. Axis x shows the input value of the machine learning model, which can be a vector containing a set of input parameters (such as the parameters listed in step 102 above). Axis y shows the output value of the machine learning model. All possible output values ​​are in the range of 0 to 1, where the two numbers 0 and 1 are the lower and upper boundaries of the output value. The two boundary numbers 0 and 1 are pre-assigned to represent two interference states, respectively. For example, the number 0 can be designated to represent an undisturbed state, and the number 1 can be designated to represent a disturbed state. The opposite representation is also possible, that is, the number 0 is used for the disturbed state and the number 1 is used for the undisturbed state. Whichever representation is used, it should be used consistently throughout the training and application of the machine learning model.

[0067] In the following, as an example, the method 100 uses the representation of the undisturbed state as a number 0 and the disturbed state as a number 1. In this case, in step 103, an output value closer to 0 indicates that the receiver is more likely or more likely to be undisturbed in the IF band in use, and vice versa, an output value closer to 1 indicates that the receiver is more likely or more likely to be disturbed in the IF band in use.

[0068] After the receiver obtains the likelihood value from the output of the machine learning model in step 103, the likelihood value can be directly used as an estimate of the interference in the receiver. In an example implementation of method 100, the receiver can also determine the interference state based on the likelihood value and a predetermined threshold. The threshold is a number between 0 and 1 and can be predetermined by the user for the receiver. Figure 3 As shown, the dotted line is located at an example threshold, for example, at 0.5. Given the threshold, when the likelihood value is greater than the threshold and the number 1 represents an interfered state, the receiver can determine that the interference state is an interfered state. Vice versa, when the likelihood value is less than or equal to the threshold and the number 0 represents an uninterrupted state, the receiver can determine that the interference state is an uninterrupted state. By using the threshold, the interference state is directly indicated, so that the interference estimation result is an indication of the interference state.

[0069] When the interference state is indicated as interfered, it can be interpreted as interference may exist in the IF band in use and may exist in the corresponding RF band in use. It can be assumed that it is unreliable for the RF band in use to be further used. If interference mitigation is performed in the reception and processing chain of the received GNSS signal, while the interference state is still indicated as interfered in the IF band in use, it can be interpreted as residual interference may exist in the IF band in use.

[0070] The method 100 may be performed continuously such that the method restarts from step 101 by receiving a further GNSS signal.

[0071] Steps 102 and 103 may be performed periodically at certain intervals, such that a set of parameters listed above in step 102 are measured and collected based on GNSS signals received in the previous interval, and estimates and indications are calculated based on the parameters collected in the previous interval. The intervals may be determined as needed or as permitted by the receiver. The intervals may depend on how much time is required to measure and collect the parameters. For example, estimates may be performed multiple times per second.

[0072] Figure 4 A flow chart of a method 200 for training a machine learning model is shown. The training method 200 is performed by a computer device. The computer device collects training data from at least a GNSS receiver in a group of GNSS receivers of the same receiver type, so that the trained machine learning model is specifically optimized for the receiver and receivers of the same receiver type. The trained machine learning model can be used by the receiver and receivers of the same receiver type to estimate interference by performing method 100 to achieve an optimal estimate of interference.

[0073] In order to train the machine learning model, a large training data set is prepared. The training data set includes interfered GNSS signals and GNSS signals that are not interfered with at multiple RF bands. For example, the GNSS signal can be different types of GNSS signals at different RF bands centered on its corresponding carrier frequency. The simulator can be used to simulate interfered GNSS signals and GNSS signals that are not interfered with. As another example, the simulator can be used to simulate different types of GNSS signals and simulate different types of interference signals. The simulated interference signal is added to a portion of the simulated GNSS signal. The rest of the simulated GNSS signal is not interfered with.

[0074] The simulated interference may include different types of interference signals, including continuous wave signals, narrowband signals, and broadband signals; the simulated interference signals may have different power levels, different frequencies, and different durations.

[0075] In the training data set, information about the GNSS signal is included together with the GNSS signal. The information may include the in-use RF band and in-use IF band of each GNSS signal, and the interference state of the signal in its in-use band. For example, when the number 0 is designated for an uninterfered state and the number 1 is designated for an interfered state, when the signal is not interfered, the number 0 may be included in the information of the GNSS signal, and when the signal is interfered, the number 1 may be included in the information of another GNSS signal.

[0076] In step 201, a receiver receives GNSS signals at a plurality of RF frequency bands, and the received signals are processed by the receiver at a plurality of corresponding IF frequency bands, wherein the received GNSS signals include interfered GNSS signals and GNSS signals that are not interfered. For example, a training data set as described above is prepared, and the receiver receives GNSS signals included in the training data set. In each time period, the same type of GNSS signals at the same RF frequency band centered on the corresponding carrier frequency are sent to the receiver to obtain measurement results of the signals in the same frequency band in use.

[0077] In step 202, the receiver records a set of parameters corresponding to the interfered GNSS signals and the GNSS signals that are not interfered with at the corresponding IF frequency bands. The set of parameters is those parameters to be used in the machine learning model to be trained and to be used to estimate interference. For example, the set of parameters can be the parameters listed in step 102 of method 100 above. In this case, when the receiver estimates interference using the trained machine learning model, the same parameters are used in method 100. In each time period, parameters are collected for the same type of GNSS signal at the same frequency band in use, and after the receiver performs steps 201 and 202 for different types of GNSS signals in the training data set, the set of parameters for multiple IF frequency bands is obtained. The set of parameters is also included in the information about the GNSS signal in the training data set, so that the information includes the frequency band in use, the interference state, and the parameters recorded at the receiver of each GNSS signal.

[0078] As an example implementation, more than one receiver in a group of GNSS receivers of the same receiver type may be used to perform steps 201 and 202 to construct a single training data set.

[0079] In step 203, the computing device trains a machine learning model based on the recorded set of parameters and the corresponding GNSS signals at the corresponding IF bands. For example, all information including the frequency band in use, the interference state, and the parameters recorded at the receiver of each GNSS signal is provided for training, wherein the recorded parameters of each GNSS signal are the input of the machine learning model, and the interference state of the corresponding GNSS signal at its IF band in use is the expected output of the machine learning model.

[0080] A single machine learning model is trained based on all information of all GNSS signals in the training data set, so that the single machine learning model can be used to estimate and indicate the interference status of the corresponding frequency band in use. In addition, since receivers of the same receiver type share similar configurations and characteristics, the single machine learning model can be used by the receiver that records the set of parameters, as well as by all receivers of the same receiver type, to achieve the best estimate of interference.

[0081] As an example implementation, the machine learning model to be trained is a logistic regression model as described in method 100.

[0082] Figure 5 An example implementation of a device including a GNSS receiving unit RECV and a processing unit PROC and an antenna or an antenna array is shown. As an example implementation, the device may be an independent electronic device such as a GNSS positioning and navigation device, or may be a mobile terminal device including RECV and PROC for GNSS related applications, or may be an electronic module embedded in a sensor device to provide GNSS related information to the sensor device, etc.

[0083] The device is configured to perform the steps performed by a GNSS receiver according to one of the implementations of method 100. The device may also be configured to perform the steps performed by a GNSS receiver according to one of the implementations of method 200.

[0084] Thus, utilizing the various implementations described above for the improved concepts, a receiver obtains an indication of the interference state in a particular frequency band in use, a machine learning model is optimally trained for the receiver and no past states are required, so that interference estimation in the receiver can be performed in a timely and lightweight manner.

[0085] Various embodiments of the improved processing concept may be implemented in the form of logic in software or hardware or a combination of both. The logic may be stored in a computer-readable or machine-readable storage medium as a set of instructions suitable for directing one or more processors of a (distributed) computer system to perform a set of steps disclosed in an embodiment of the improved processing concept. The logic may form part of a computer program product suitable for directing an information processing device to automatically perform a set of steps disclosed in an embodiment of the improved processing concept.

[0086] Accordingly, the specification and drawings are to be regarded in an illustrative rather than a restrictive sense. It will, however, be evident that various modifications and changes may be made thereto without departing from the scope of the present invention as set forth in the claims.

Claims

1. A method for estimating interference in a Global Navigation Satellite System (GNSS) receiver, the method being performed in the receiver and comprising: receiving GNSS signals at a radio frequency band and processing the received signals at an intermediate frequency band; collecting a set of parameters at the receiver based on the received GNSS signals; as well as A likelihood value is obtained using a machine learning model trained for the receiver, wherein the set of parameters is an input to the machine learning model, the likelihood value is an output of the machine learning model, and the likelihood value is a number between 0 and 1 and corresponds to the likelihood of the receiver being interfered in the intermediate frequency band.

2. The method according to claim 1, wherein: The machine learning model is a logistic regression model.

3. The method according to claim 1, wherein: The set of parameters includes: the skewness of the radio frequency spectrum at the radio frequency band; the gain of the radio frequency amplifier in the receiver; the skewness of the IF spectrum at the IF band; The variance of the IF spectrum at the IF band; the variance of the IF histogram at the IF band; an average value of a collapsed modulus-1 kHz spectrum at the intermediate frequency band, wherein the collapsed modulus-1 kHz spectrum is selected from frequency components having modulus-1 kHz values ​​in a spectrum around the intermediate frequency; and It tracks the number of GNSS signals that are aborted by the receiver due to high correlation with noise.

4. The method according to any one of claims 1 to 3, wherein: The method also includes applying interference mitigation to the received GNSS signals prior to collecting the set of parameters.

5. The method according to any one of claims 1 to 3, further comprising: Determine a threshold value, wherein the threshold value is a number between 0 and 1; When the possibility value is greater than the threshold, determining that the interference state is an interfered state; When the possibility value is less than or equal to the threshold, the interference state is determined to be a non-interference state.

6. A method of training a machine learning model by a computing device for estimating interference in a Global Navigation Satellite System (GNSS) receiver, the method comprising: receiving GNSS signals at a plurality of radio frequency bands by the receiver, and processing the received signals at a plurality of corresponding intermediate frequency bands by the receiver, wherein the received GNSS signals include interfered GNSS signals and non-interfered GNSS signals; recording, by the receiver, a set of parameters corresponding to an interfered GNSS signal and an undisturbed GNSS signal at a corresponding MF band; and The machine learning model is trained by the computing device based on the recorded set of parameters and corresponding GNSS signals at the corresponding medium frequency band.

7. The method according to claim 6, wherein: Recording a set of parameters includes recording, by the receiver, the set of parameters reflecting information on whether the received GNSS signals are interfered or not interfered within the corresponding MF band.

8. The method according to claim 6, wherein: The set of parameters includes: skewness of the radio frequency spectrum at the plurality of radio frequency bands; the gain of the radio frequency amplifier in the receiver; skewness of the IF spectrum at the plurality of IF bands; variance of the intermediate frequency spectrum at the plurality of intermediate frequency bands; variance of the intermediate frequency histogram at the plurality of intermediate frequency bands; an average value of collapsed modulus-1 kHz spectra at the plurality of intermediate frequency bands, wherein the collapsed modulus-1 kHz spectrum is selected from frequency components having modulus-1 kHz values ​​in the spectrum at the intermediate frequency; and The number of GNSS channels whose tracking is aborted by the receiver due to high correlation with noise.

9. The method according to any one of claims 6 to 8, wherein: Training the machine learning model includes: Setting an interference state of the GNSS signal at the corresponding intermediate frequency, wherein the interference state is a state in which the GNSS signal is interfered with or a state in which the GNSS signal is not interfered with; and The machine learning model is trained based on the recorded parameters and the corresponding interference state of the GNSS signal at the corresponding intermediate frequency.

10. The method according to any one of claims 6 to 8, further comprising: simulating an interference signal by the computing device; as well as applying the simulated jamming signal to at least a portion of the GNSS signal to be received by the receiver; Wherein, the simulated interference signal: There are different types of signals including continuous wave signals, narrowband signals and broadband signals; Available in different power levels; have different frequencies; and With different durations.

11. A device for estimating interference in a global navigation satellite system (GNSS) receiver, the device comprising a global navigation satellite system (GNSS) receiving unit and a processing unit, the device being configured to: receiving GNSS signals at a radio frequency band and processing the received signals at an intermediate frequency band; collecting a set of parameters at the device; as well as A likelihood value is obtained using a machine learning model trained for the device, wherein the set of parameters is an input to the machine learning model, the likelihood value is an output of the machine learning model, and the likelihood value is a number between 0 and 1 and corresponds to the likelihood of the device being interfered with in the mid-frequency band.

12. The device according to claim 11, wherein The machine learning model is a logistic regression model.

13. The device according to claim 11, wherein: The set of parameters includes: the skewness of the radio frequency spectrum at the radio frequency band; the gain of the radio frequency amplifier at the device; the skewness of the IF spectrum at the IF band; The variance of the IF spectrum at the IF band; the variance of the IF histogram at the IF band; an average value of a collapsed modulus-1 kHz spectrum at the intermediate frequency band, wherein the collapsed modulus-1 kHz spectrum is selected from frequency components having modulus-1 kHz values ​​in the spectrum at the intermediate frequency; and It tracks the number of GNSS signals that are aborted by the device due to high correlation with noise.

14. The apparatus according to any one of claims 11 to 13, wherein: The device is also configured to apply interference mitigation to the received GNSS signals prior to collecting the set of parameters.

15. The device according to any one of claims 11 to 13, further configured to: receiving further GNSS signals at a plurality of radio frequency bands and processing the further GNSS signals at a plurality of intermediate frequency bands, wherein: The other GNSS signals include interfered GNSS signals and undisturbed GNSS signals; as well as Record another set of parameters for training the machine learning model, wherein the another set of parameters corresponds to the interfered and undisturbed further GNSS signals at the corresponding mid-frequency band.

Images