Industrial electronic control system radio frequency signal interference regulation method and system

Abstract

The application discloses an industrial electronic control system radio frequency signal anti-interference regulation method and system, and belongs to the technical field of industrial electronic control electromagnetic anti-interference. A cross-domain coupling model is constructed by acquiring the nonlinear rectification offset and timing jitter characteristics generated by the analog front-end circuit under the action of the radio frequency field. A cooperative regulation instruction containing analog front-end bias compensation parameters and digital control timing reconstruction parameters is outputted and generated. The bias compensation parameters are used to offset the direct current and low frequency error caused by the rectification offset. The timing reconstruction parameters are used to adjust the sampling time or communication time slot to avoid the timing sensitive window. The cooperative regulation instruction is sent to the hardware execution unit, so that the analog front-end circuit and the digital control link maintain the preset control precision threshold. The application can simultaneously process the analog domain rectification offset and the digital domain timing jitter, suppress the additional error generated by the cross coupling of the two, and improve the electromagnetic anti-interference performance of the control system in the complex radio frequency industrial scene.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic interference suppression technology for industrial electronic control, and specifically to a method and system for radio frequency signal interference suppression and control in industrial electronic control systems. Background Technology

[0002] Industrial production sites are filled with numerous electromagnetic radiation sources, such as walkie-talkies, wireless base stations, frequency converters, and radio frequency testing equipment. The analog sampling front-end and clock digital link of industrial electronic control systems are highly susceptible to interference from spatial radio frequency radiation and cable-conducted coupling. When radio frequency energy intrudes into the controller hardware, it doesn't just manifest as ordinary superimposed noise; it also undergoes radio frequency rectification through the nonlinear effect of the chip's PN junction, converting high-frequency radio frequency energy into DC offset error. Simultaneously, radio frequency radiation interferes with the clock distribution network, inducing clock edge jitter, ultimately causing sampled value drift and closed-loop control output deviation. In severe cases, this can lead to actuator malfunctions, production shutdowns, and other safety accidents. Traditional industrial electronic control systems rely primarily on static hardware protection such as shielding, grounding, and LC filtering, combined with software anti-interference measures like fixed-parameter digital filtering and multiple sampling averaging. However, these solutions are all preset static protections and cannot dynamically adjust protection strategies according to the time-varying radio frequency interference environment. In complex industrial scenarios where interference frequencies and field strengths dynamically change, the anti-interference compensation effect is significantly limited.

[0003] Numerous academic papers, both domestically and internationally, have analyzed the mechanisms by which radio frequency (RF) interference causes analog front-end rectification offset and digital clock jitter. ADI's official technical note MT-096, "RFI Rectification Principle," systematically explains the physical mechanism of DC offset generated by RF rectification in the operational amplifier input stage. It clarifies that high-frequency out-of-band RF interference enters the analog device ports via radiation or cable coupling, generating DC offset through transistor nonlinear detection, which is the core cause of sampling drift under RF interference in industrial sampling circuits. A paper published in the *Journal of Electronics and Information Technology*, titled "Second-Order Intermodulation Low-Frequency Blocking Effect and Mechanism of Electromagnetic Radiation from Communication Radio Stations," demonstrates through irradiation experiments that second- and third-order nonlinear intermodulation products of circuits under RF radiation are converted into low-frequency baseband interference, directly degrading the output accuracy of the acquisition link. Most existing publicly available academic research analyzes analog link rectification offset and digital link clock jitter as two independent interference phenomena, paying little attention to the objectively existing cross-coupling and transmission relationship between the two types of errors. At the same time, most existing research is still at the level of offline mechanism verification in the laboratory, which requires testing on a dedicated irradiation test bench. It lacks real-time quantitative sensing methods for equipment under normal online operating conditions, making it difficult to simultaneously obtain rectification offset and RF jitter characteristics without interrupting the operation of the controller. The corresponding theoretical results are difficult to directly transform into online anti-interference control algorithms that can be used in engineering. They can only complete the explanation of the phenomena and cannot form a real-time control algorithm that can be directly deployed inside the industrial controller.

[0004] Existing domestic patents have proposed corresponding technical solutions for electromagnetic suppression and digital-analog nonlinearity cancellation of control boards. The invention patent with publication number CN120546455A, "A Method for Suppressing Electromagnetic Interference of Control Boards", discloses the collection of board operating status and electromagnetic environment data, and the post-processing of the collected signals through dynamic digital filtering to weaken the impact of electromagnetic interference on the output signals of the control board. It can suppress interference signals by relying on software filtering and reduce data fluctuations caused by electromagnetic noise. However, this solution only performs post-processing filtering correction on the already generated sampling results, and cannot identify the DC offset component caused by the radio frequency rectification effect. It also does not consider the timing jitter problem caused by radio frequency radiation intruding into the digital clock network. The invention patent CN119324752A, entitled "A Nonlinear Analog-Digital Joint Cancellation Method, System, Device and Medium," discloses the acquisition of analog and digital domain output data and the construction of a nonlinear cancellation model to compensate for signal distortion. Although it involves signal processing in both analog and digital domains, this solution cancels the nonlinear distortion of the signal transmission path itself, rather than the rectification offset and RF-induced clock jitter induced by external space RF radiation. It does not design a sensing link for online sensing of RF interference characteristics, nor does it have the technical approach of constructing a cross-domain coupling model based on the joint distribution of the two types of interference errors. Neither of the two patent solutions can simultaneously sense the analog-side DC offset and digital-side timing jitter induced by the RF field, and can only perform single-dimensional signal correction. When RF simultaneously acts on the analog-digital hardware link in the industrial field, the comprehensive error generated by the coupling superposition cannot be suppressed in a targeted manner, and the control system still faces the risk of exceeding the control accuracy limit under strong RF conditions.

[0005] Existing academic research is mostly limited to offline mechanism analysis in the laboratory, lacking an engineering-based online sensing and collaborative compensation implementation path. Existing related patented technologies handle board electromagnetic noise or transmission link nonlinear distortion independently. The two technical solutions are disconnected from each other, ignoring the analog rectification offset, digital clock jitter, and cross-domain coupling effect caused by radio frequency radiation. It is impossible to quantify the two types of interference effects simultaneously, nor can it construct a unified cross-domain coupling model based on the joint distribution of the two to output collaborative control commands. The residual coupling error after compensation is difficult to eliminate, making it difficult to meet the practical needs of high-precision closed-loop operation of control systems in complex industrial radio frequency environments. Summary of the Invention

[0006] Based on the aforementioned technical problems, this application discloses a method and system for interference suppression and control of radio frequency signals in industrial electronic control systems; the method for interference suppression and control of radio frequency signals in industrial electronic control systems includes:

[0007] To obtain the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of radio frequency field, and the timing jitter characteristics of the digital control link under the current electromagnetic environment;

[0008] Based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics, a cross-domain coupling model is constructed to characterize the contribution of radio frequency interference to the overall error of the control system.

[0009] Based on the output of the cross-domain coupling model, a coordinated control command is generated, which includes analog front-end bias compensation parameters and digital control timing reconstruction parameters. The analog front-end bias compensation parameters are used to offset the DC and low-frequency error components introduced by the nonlinear rectification offset, and the digital control timing reconstruction parameters are used to adjust the sampling time or communication time slot to avoid the timing-sensitive window caused by radio frequency interference.

[0010] The coordinated control command is sent to the corresponding hardware execution unit so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

[0011] Preferably, the acquisition of the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of a radio frequency field includes:

[0012] An orthogonal dual-frequency radio frequency detection signal is injected into the input stage of the analog front-end circuit, the frequency of which is located in the out-of-band frequency band of the industrial electronic control system;

[0013] The baseband response signal at the output of the analog front-end circuit is acquired, and the intermodulation components are extracted from the baseband response signal to obtain the amplitude of the second-order intermodulation product and the amplitude of the third-order intermodulation product.

[0014] The nonlinear rectification offset is calculated based on the amplitude of the second-order intermodulation product, the amplitude of the third-order intermodulation product, and the injection power of the orthogonal dual-frequency radio frequency detection signal.

[0015] Preferably, the nonlinear rectification offset is calculated based on the amplitudes of the second-order intermodulation products, the amplitudes of the third-order intermodulation products, and the injection power of the orthogonal dual-frequency radio frequency detection signal, using the following formula: ;in, This refers to the nonlinear rectification offset. The injection power of the orthogonal dual-frequency radio frequency detection signal. For load impedance, and These are the second-order and third-order nonlinear coefficients of the analog front-end circuit, respectively. and These are the amplitudes of the second-order intermodulation product and the amplitudes of the third-order intermodulation product, respectively. The frequency of the orthogonal dual-frequency radio frequency detection signal , Related environmental coupling correction factors.

[0016] Preferably, the acquisition of timing jitter characteristics of the digital control link under the current electromagnetic environment includes:

[0017] Multiple jitter sampling nodes are set in the clock distribution network of the digital control link, and the edge offset sequence of each jitter sampling node is collected in multiple consecutive clock cycles.

[0018] The edge offset sequence is decomposed in the frequency domain to extract periodic jitter components and random jitter components;

[0019] Based on the spectral peak distribution of the periodic jitter components, the radio frequency induced jitter components related to the frequency of the interference source in the radio frequency field are identified, and the amplitude and phase information of the radio frequency induced jitter components are used as the timing jitter features.

[0020] Preferably, the radio frequency induced jitter component related to the frequency of the interference source in the radio frequency field is identified based on the spectral peak distribution of the periodic jitter component, using the following formula:

[0021]

[0022] in, For frequency The power spectral density of the radio frequency sensed jitter component at that location. The first edge offset sequence Each sample value, The number of sampling points. For window functions, The source of interference in the radio frequency field at frequency Power spectral density estimation at [location] The background noise power spectral density of the digital control link is given. is the regularization constant.

[0023] Preferably, the construction of a cross-domain coupling model characterizing the contribution of radio frequency interference to the overall error of the control system based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics includes:

[0024] A multidimensional probability space is constructed with the nonlinear rectification offset and the timing jitter characteristics as input variables, and the joint probability density function of the two is calculated in the multidimensional probability space.

[0025] Based on the joint probability density function, the cross-domain coupling model is established, and the output of the cross-domain coupling model is the statistical expectation value of the comprehensive error of the control system.

[0026] The model parameters of the cross-domain coupling model are updated online by minimizing the deviation between the statistical expected value and the preset control accuracy threshold.

[0027] Preferably, the output of the cross-domain coupling model is the statistical expectation value of the integrated error of the control system, and the formula for the statistical expectation value is:

[0028]

[0029] in, This is the statistical expectation value of the overall error of the control system; The instantaneous comprehensive error function is determined by the analog front-end error propagation relationship, the digital link jitter propagation relationship, and the analog-digital cross-coupling relationship. It is the joint probability density function of nonlinear rectified offset and RF induction jitter component; This is the integration domain.

[0030] A radio frequency signal interference suppression and control system for industrial electronic control systems includes a multi-source sensing module, a cross-domain coupling modeling module, a collaborative control decision-making module, and a hardware execution module, specifically:

[0031] The multi-source sensing module acquires the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of the radio frequency field, as well as the timing jitter characteristics of the digital control link in the current electromagnetic environment.

[0032] The cross-domain coupling modeling module constructs a cross-domain coupling model that characterizes the contribution of radio frequency interference to the overall error of the control system based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics.

[0033] The collaborative control decision module generates a collaborative control instruction that includes analog front-end bias compensation parameters and digital control timing reconstruction parameters based on the output of the cross-domain coupling model.

[0034] The hardware execution module sends the coordinated control command to the corresponding hardware execution unit, so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

[0035] An electronic device, comprising:

[0036] At least one processor; and,

[0037] A memory communicatively connected to the at least one processor; wherein,

[0038] The memory stores instructions that can be executed by the at least one processor, which enables the at least one processor to perform the radio frequency signal anti-interference control method for industrial electronic control systems.

[0039] A computer-readable storage medium storing computer instructions for causing a computer to execute the described method for radio frequency signal interference suppression and control in industrial electronic control systems.

[0040] Compared with the prior art, the technical solution of this application has the following technical effects:

[0041] This invention acquires in real time the nonlinear rectification offset caused by radio frequency interference and the timing jitter characteristics caused by radio frequency induction in industrial electronic control systems. Based on the online acquisition of the two types of interference-related characteristic information through the hardware link, it can conduct offline irradiation testing without interrupting the normal operation of the equipment. It realizes the synchronous perception of two types of interference elements on the analog and digital sides, and provides a reliable input basis for subsequent model construction and control decisions.

[0042] This invention constructs a cross-domain coupling model based on the joint distribution of two types of interference characteristics, which fully covers the transmission of analog link errors, the transmission of digital link jitter, and the cross-coupling relationship between analog and digital. By using statistical expected values ​​to characterize the comprehensive error of the control system caused by radio frequency interference, it can fully reflect the true error state of the system under the interaction of the two types of interference, and improve the completeness of the characterization of the degree of influence of radio frequency interference.

[0043] This invention generates coordinated control commands based on the output of a cross-domain coupling model. The control commands simultaneously include analog front-end bias compensation parameters and digital control timing reconstruction parameters, which can be applied synchronously to the analog circuit and the digital control link to respectively address the errors caused by rectification offset and the adverse effects caused by timing jitter, thereby realizing coordinated anti-disturbance control operations in the analog and digital domains.

[0044] This invention sends coordinated control commands to the hardware execution unit to perform bias compensation on the analog front-end circuit, offsetting various error components introduced by nonlinear rectification offset. At the same time, it performs timing adjustment on the sampling time of the digital link and the communication time slot, avoiding the timing sensitive window corresponding to radio frequency interference, and weakening the adverse effects of radio frequency interference on the control system from the source of interference.

[0045] This invention, through the combined action of simulated bias compensation and timing reconstruction, works on the hardware link of the control system to effectively reduce the impact of various errors induced by radio frequency interference on the closed-loop control process, ensuring that the industrial electronic control system maintains the predetermined control accuracy under the condition of radio frequency interference, and enhancing the working stability of the control system in the face of radio frequency interference.

[0046] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the preferred embodiments of this application are described in detail below with reference to the accompanying drawings.

[0047] The above and other objects, advantages and features of this application will become more apparent to those skilled in the art from the following detailed description of specific embodiments in conjunction with the accompanying drawings. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0049] Based on the description of the figures and their corresponding technical content in the document, the titles of the figures are as follows:

[0050] Figure 1 A flowchart illustrating the overall process of anti-interference control methods for industrial electronic control systems under radio frequency interference.

[0051] Figure 2 A flowchart of the signal processing procedure for solving the nonlinear rectified offset of orthogonal dual-frequency detection signals;

[0052] Figure 3 A flowchart illustrating the complete process of clock jitter multi-node sampling and RF sensing jitter feature extraction;

[0053] Figure 4 This is a block diagram illustrating the principle of error propagation and feedback coupling in an analog-digital cross-domain coupled model.

[0054] Figure 5 A flowchart illustrating the closed-loop execution process of simulated bias compensation and digital timing reconstruction coordinated control;

[0055] Figure 6 A time-series comparison curve of system integrated error under multi-source aliasing radio frequency interference conditions;

[0056] Figure 7 A bar chart comparing the RMSE performance of each test model under different RF field strength ranges;

[0057] Figure 8This is a scatter plot-curve combination of the system's comprehensive error time series under multiple control modes and parameter comparison conditions.

[0058] Figure 9 A combination of subplot curves showing the evolution of key intermediate physical quantities and model loss function under multiple disturbance conditions;

[0059] Figure 10 A heatmap of the steady-state average error distribution of the system under multiple hardware parameter scanning conditions;

[0060] Figure 11 Diagram of analog-digital cross-domain collaborative anti-interference closed-loop control architecture for industrial electronic control systems under radio frequency interference environment;

[0061] Figure 12 A schematic diagram of the hardware structure of an electronic device for running cross-domain cooperative anti-interference program instructions for radio frequency interference. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. In the following description, specific details such as specific configurations and components are provided merely to help fully understand the embodiments of this application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. In addition, for clarity and brevity, descriptions of known functions and structures are omitted in the embodiments.

[0063] It should be understood that the phrase "an embodiment" or "this embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "an embodiment" or "this embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0064] Furthermore, reference numerals and / or letters may be repeated in different examples within this application. Such repetition is for the purpose of simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or settings discussed.

[0065] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" in this article describes another type of relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the related objects before and after it are in an "or" relationship.

[0066] In this article, the term "at least one" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, "at least one of A and B" can mean: A exists alone, A and B exist simultaneously, or B exists alone.

[0067] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion.

[0068] Example 1

[0069] This embodiment mainly describes a method for anti-interference control of radio frequency signals in industrial electronic control systems, such as... Figure 1 As shown, it includes:

[0070] To obtain the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of radio frequency field, and the timing jitter characteristics of the digital control link under the current electromagnetic environment;

[0071] Based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics, a cross-domain coupling model is constructed to characterize the contribution of radio frequency interference to the overall error of the control system.

[0072] Based on the output of the cross-domain coupling model, a coordinated control command is generated, which includes analog front-end bias compensation parameters and digital control timing reconstruction parameters. The analog front-end bias compensation parameters are used to offset the DC and low-frequency error components introduced by the nonlinear rectification offset, and the digital control timing reconstruction parameters are used to adjust the sampling time or communication time slot to avoid the timing-sensitive window caused by radio frequency interference.

[0073] The coordinated control command is sent to the corresponding hardware execution unit so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

[0074] An orthogonal dual-frequency radio frequency detection signal is injected into the input stage of the analog front-end circuit, the frequency of which is located in the out-of-band frequency band of the industrial electronic control system.

[0075] like Figure 2 As shown, the orthogonal dual-frequency radio frequency detection signal is generated by the radio frequency signal generation unit, and it contains two frequency components, denoted as the first detection frequency. Second detection frequency The first detection frequency With the second detection frequency Frequency difference between The following conditions must be met: ;in, The bandwidth of the useful signal in the industrial electronic control system. The frequency difference. The selection of the second-order intermodulation product makes Falling within the useful signal bandwidth or near-band region, it can effectively excite the nonlinear rectification response of the analog front-end circuit under the action of the radio frequency field. The first detection frequency Second detection frequency All are located outside the useful signal bandwidth, specifically satisfying: ;in, The highest frequency component of the useful signal. To protect the bandwidth, the value is set between 1.5 and 3 times the bandwidth of the useful signal to ensure that the detection signal does not alias with the useful signal in the frequency domain.

[0076] The injection power of the orthogonal dual-frequency radio frequency detection signal It needs to be set according to the linear dynamic range of the analog front-end circuit. Specifically, the injected power The following constraints must be met: ,in, The input power corresponding to the 1dB compression point of the analog front-end circuit. This is a power margin, ranging from 6dB to 12dB. This constraint ensures that while the probe signal excites the nonlinear rectification effect, it does not cause the analog front-end circuit to enter a deep saturation region, thus guaranteeing that the extracted intermodulation components can accurately reflect the nonlinear characteristics of the circuit near its normal operating point.

[0077] The orthogonal dual-frequency radio frequency detection signal is injected using orthogonal modulation, meaning that the first and second detection frequency components maintain a 90-degree phase difference. Its time-domain formula is: ;

[0078] in, To detect the amplitude of the signal, and the injected power The relationship is: , For load impedance, The characteristic impedance of the system is given. The purpose of using quadrature modulation is to give the second-order and third-order intermodulation products different phase characteristics in the frequency domain, which facilitates subsequent signal separation and extraction. At the same time, quadrature injection can reduce the transient disturbance amplitude of the probe signal to the normal operation of the analog front-end circuit.

[0079] The baseband response signal at the output of the analog front-end circuit is acquired, and the intermodulation components of the baseband response signal are extracted to obtain the amplitude of the second-order intermodulation product and the amplitude of the third-order intermodulation product.

[0080] The baseband response signal is acquired via a high-speed analog-to-digital converter (ADC), and the sampling rate of the high-speed ADC is... satisfy: ;in, The highest intermodulation product frequency that needs to be extracted. The bandwidth of the intermodulation product. In a typical industrial electronic control system, the frequency of the second-order intermodulation product is... The frequency of the third-order intermodulation product is Therefore, the sampling rate must satisfy: ;

[0081] Acquired baseband response signal It includes the useful signal component, the residual fundamental component of the probe signal, second-order intermodulation components, third-order intermodulation components, and circuit background noise. To extract the intermodulation components, the baseband response signal is... Perform the following processing:

[0082] Band-stop filtering removes useful signal components. A digital band-stop filter is designed so that its stopband covers the frequency band of the useful signal. The filter's stopband attenuation should be no less than 40dB, and its passband ripple should not exceed 0.5dB. The filtered signal is denoted as... .

[0083] Coherent demodulation extracts second-order intermodulation products using a local reference signal. right Perform coherent demodulation:

[0084]

[0085] For in-phase components and orthogonal components Perform low-pass filtering separately, with the cutoff frequency set to... The DC component is obtained. and Second-order intermodulation product amplitude Calculated by the following formula: ;

[0086] Coherent demodulation extracts third-order intermodulation products using a similar method, with a local reference signal. Coherent demodulation was performed to obtain the amplitude of the third-order intermodulation product. : ;

[0087] In the above coherent demodulation process, the equivalent integration time of the low-pass filter Must meet: ;in, This is the signal-to-noise ratio gain factor, ranging from 10 to 50, to ensure the accuracy of intermodulation product amplitude estimation. A longer integration time results in stronger noise suppression, but a correspondingly slower response speed; a trade-off must be made based on the rate of change in the actual RF environment.

[0088] The nonlinear rectification offset is calculated based on the amplitude of the second-order intermodulation product, the amplitude of the third-order intermodulation product, and the injection power of the orthogonal dual-frequency radio frequency detection signal.

[0089] The nonlinear rectification offset The physical essence is that when semiconductor devices in the analog front-end circuit (including the emitter junction of bipolar transistors, the gate-source junction of field-effect transistors, and the PN junction of the differential input pair of operational amplifiers) are excited by radio frequency signals, due to the nonlinearity of their current-voltage characteristics (i.e., the relationship between current and voltage does not satisfy a strict linear ratio), the envelope information of the radio frequency signal is "rectified" or "demodulated" to the baseband frequency band, and an additional offset voltage related to the strength of the radio frequency signal is superimposed on the DC operating point of the circuit.

[0090] The nonlinear rectification offset Calculated using the following formula: ;

[0091] Injected power term This reflects the injected probe signal at the load impedance The equivalent voltage squared is generated. The RF rectification effect is essentially a second-order nonlinear effect; the resulting DC offset is proportional to the square of the input signal amplitude, and therefore related to the injected power. The relationship is quadratic. Load impedance To simulate the equivalent load impedance of the input stage of the front-end circuit, its value is determined by the circuit topology, typically ranging from 50Ω to 10kΩ. Second-order nonlinear coefficient. With third-order nonlinear coefficients Describe the degree of nonlinearity in the input-output characteristics of the analog front-end circuit. Expand the input-output transfer characteristics of the analog front-end circuit using Taylor series: ;in, For linear gain, These are second-order nonlinear coefficients. These are third-order nonlinear coefficients. These two coefficients are obtained through pre-calibration of the analog front-end circuit: in a shielded environment, a sinusoidal signal of known amplitude is input to the circuit, the second and third harmonic components at the output are measured, and then calculated. and For a typical industrial operational amplifier, The order of magnitude is , The order of magnitude is .

[0092] Second-order intermodulation product amplitude and the amplitude of third-order intermodulation products These two quantities are obtained from the aforementioned coherent demodulation process and respectively reflect the intensity of second-order and third-order nonlinear distortion of the analog front-end circuit under the action of the radio frequency field. It primarily characterizes the ability of an RF signal envelope to be rectified into DC / low-frequency components. This characterizes the contribution of higher-order nonlinear effects to the rectified offset. Environmental coupling correction factor. This factor is used to correct for the impact of multipath reflection, cavity resonance, and PCB trace parasitic parameters on RF energy coupling efficiency in actual industrial electromagnetic environments. Its calculation method is as follows: ;in, and These are analog front-end circuits at different frequencies. and The input impedance at that point, The characteristic impedance of the system, In frequency The reflection coefficient at that location, For the first The weighting coefficients of each resonant mode This represents the total number of resonant modes considered. This correction factor, obtained by fitting after frequency sweep measurements in a real-world deployment environment, significantly improves the accuracy of nonlinear rectified offset calculations.

[0093] Multiple jitter sampling nodes are set up in the clock distribution network of the digital control link, and the edge offset sequence of each jitter sampling node is collected over multiple consecutive clock cycles, such as... Figure 3 As shown;

[0094] The digital control link includes a clock generator, a clock buffer, clock distribution wiring, and a clock receiver (including the sampling clock input of the analog-to-digital converter, the system clock input of the microprocessor, and the baud rate clock input of the communication interface). The jitter sampling node is located in the following positions:

[0095] The first sampling node is located at the output of the clock generator and is used to capture the inherent jitter of the clock source itself (including random jitter caused by thermal noise and periodic jitter caused by oscillator phase noise).

[0096] The second sampling node, located at the output of the clock buffer, is used to capture additional jitter introduced by the clock buffer under radio frequency interference. The transistors inside the clock buffer also generate nonlinear rectification effects under the influence of radio frequency fields, causing their switching thresholds to shift, thus introducing additional timing jitter.

[0097] The third sampling node is located at the end of the clock distribution trace (i.e., the clock receiver) and is used to capture the final jitter state of the clock signal after being subjected to spatial radiation coupling of the radio frequency field and conductive coupling of the power / ground line during transmission.

[0098] Each jitter sampling node acquires the actual arrival time of the clock edge via a Time-to-Digital Converter (TDC). The time resolution of the TDC... Must meet: ,in, For clock cycles, This is the resolution factor, with a value ranging from 100 to 1000. In industrial electronic control systems, clock frequencies are typically between 10MHz and 200MHz, corresponding to clock periods of 5ns to 100ns. Therefore, the time resolution of the time-to-digital converter needs to be between 5ps and 100ps.

[0099] The edge offset sequence Defined as the actual clock edge arrival time Arrival time of the ideal clock edge The difference between them: ;

[0100] in, This represents the total number of clock cycles collected continuously. Must meet: ,in, For clock frequency, The lowest jitter frequency component that needs to be detected. This is the frequency resolution coefficient, ranging from 4 to 16. The acquired edge offset sequence. It also contains multiple jitter components: intrinsic random jitter, jitter caused by power supply noise, long-term jitter caused by thermal drift, and periodic jitter caused by radio frequency interference.

[0101] The edge offset sequence is decomposed in the frequency domain to extract periodic jitter components and random jitter components.

[0102] The frequency domain decomposition is achieved through the following steps:

[0103] Detrending processing is performed on the original edge offset sequence. By removing the linear trend and DC component, a detrended sequence is obtained. : ;in, The mean of the sequence. The slope of the linear trend is obtained by fitting using the least squares method: ;in, This is the sequence center index.

[0104] Windowed Fourier transform applies a window function to the detrended sequence. And perform a discrete Fourier transform: The window function The formula for using KaiserWindow is as follows: ;in, For zero-order modified Bessel functions, This is a window shape parameter, with a value range of 4 to 8. A larger Kaiser window results in stronger sidelobe suppression but also a wider main lobe. The reason for choosing the Kaiser window is that its sidelobe attenuation can be reduced by... It is flexible and suitable for detecting weak periodic jitter components in strong background noise.

[0105] Power spectral density estimation is separated from jitter components, and power spectral density is calculated: ;in, This is the energy normalization factor for the window function.

[0106] In power spectral density In this spectrum, periodic jitter components appear as discrete spectral peaks, while random jitter components appear as a broadband continuous spectral base. A peak detection threshold is set... Separate: ;in, and These represent the spectral mean and standard deviation of the power spectral density, respectively. This is the peak detection sensitivity coefficient, ranging from 3 to 6. It satisfies... frequency points The components marked as periodic jitter are classified as random jitter components, while the components at other frequencies are classified as random jitter components.

[0107] Based on the spectral peak distribution of the periodic jitter components, the radio frequency induced jitter components related to the frequency of the interference source in the radio frequency field are identified, and the amplitude and phase information of the radio frequency induced jitter components are used as the timing jitter features.

[0108] Timing jitter caused by radio frequency interference has a characteristic frequency directly related to the frequency of the interference source. Specifically, radio frequency fields introduce periodic jitter into digital clock signals through the following pathways:

[0109] The radio frequency envelope detection effect occurs when the radio frequency signal is rectified by the nonlinear input stage of the clock buffer. The envelope of the radio frequency signal (such as the amplitude modulation frequency) is directly superimposed on the threshold voltage of the clock signal, causing the clock edge to have a periodic offset on the time axis that is in sync with the envelope frequency.

[0110] Power / ground wires are elastically coupled, and radio frequency interference is coupled into the power distribution network, generating periodic ripples on the power supply voltage that are related to the radio frequency or its harmonics. These ripples are converted into clock jitter by the limited power supply rejection ratio (PSRR) of the clock generator and buffer.

[0111] Substrate coupling occurs within an integrated circuit, where radio frequency energy is coupled to the sensitive nodes of the clock circuit through the substrate, introducing jitter components related to the radio frequency.

[0112] The radio frequency sensing jitter component Calculated using the following formula: Fourier transform term The squared modulus of the windowed discrete Fourier transform of the edge offset sequence reflects the frequency shift of the edge offset. The energy concentration at that location. For clock frequency, The number of sampling points. The aforementioned Kaiser window function is used for radio frequency field power spectral density estimation. The frequency data obtained through radio frequency environment sensing antennas or spectrum monitoring modules reflects the current frequency in the electromagnetic environment. The intensity of radio frequency interference at that location. The method of acquisition is as follows: Environmental radio frequency signals are received through a broadband antenna, and power spectrum estimation is performed after down-conversion and analog-to-digital conversion. Background noise power spectral density. It reflects the inherent timing jitter level of the digital control link in an environment without radio frequency interference. It is obtained by measuring in advance in an electromagnetically shielded room and stored in the system's non-volatile memory. The typical shape is: present in the low frequency range The flicker noise characteristics exhibit flat white noise at high frequencies. Regularization constant. The value is taken as the minimum value of the background noise power spectral density. Doubled Double, used to prevent in Numerical overflow occurs at extremely low frequency points. The ratio in the formula... It serves as a radio frequency interference weighting factor. At frequency points where the radio frequency interference intensity is much higher than the noise floor, this ratio is relatively large, making... Highlights jitter components affected by radio frequency (RF); at frequency points where the RF interference intensity is close to or below the noise floor, this ratio approaches zero, thereby suppressing non-RF-related jitter components.

[0113] In addition to power spectral density, the radio frequency inductive jitter component also includes phase information. Phase information Calculated by the following formula: ;

[0114] The phase information This is used to accurately locate the starting position of the timing-sensitive window during subsequent digital control timing reconstruction, so that the time slot adjustment can accurately avoid the moment when jitter peaks occur.

[0115] In this embodiment, the cross-domain coupling model is constructed based on the following physical mechanism: when a strong radio frequency field acts on the analog front-end circuit, the nonlinear characteristics of the semiconductor PN junction rectify the radio frequency energy into baseband DC and low-frequency error components (i.e., the nonlinear rectification offset); simultaneously, radio frequency interference causes timing jitter in the digital control link through power / ground coupling or spatial radiation. The inventors have discovered that these two effects are not independent but rather have a significant coupling amplification relationship: the baseband error generated by rectification changes the threshold voltage of the digital circuit, thus exacerbating the sensitivity to timing jitter; conversely, the sampling deviation caused by timing jitter causes the rectification error to be sampled non-uniformly, further deteriorating control accuracy. The cross-domain coupling model is constructed precisely to quantify the contribution of this coupling relationship to the overall system error, such as... Figure 4 As shown;

[0116] The microscopic physical processes involved in the aforementioned coupling amplification relationship can be further refined into the following three levels:

[0117] The modulation effect of rectification offset on timing jitter. DC offset voltage of the operational amplifier output stage in the analog front-end circuit. The propagation delay of digital logic gates is conducted to the power rails of digital circuits through common-mode feedback networks or power distribution networks. With supply voltage The relationship between them can be approximated as: ;in, Nominal supply voltage The propagation delay below, This is a process-related index; for CMOS circuits using process nodes of 65nm and below, The value range is from 1.2 to 1.8. When When the value is not zero, the propagation delay shifts, and this shift directly modulates the position of the clock edge, resulting in additional timing jitter. The sensitivity of this modulation effect... Defined as: Sensitivity at a nominal supply voltage of 3.3V and a nominal propagation delay of 2ns. At that time, the calculated value is approximately That is, every 1mV of rectification offset will introduce an additional timing jitter of about 0.91ps.

[0118] Timing jitter causes sampling distortion due to rectification errors. When the analog-to-digital converter in the digital control link samples with a jittered clock, the uncertainty in the sampling timing will lead to non-uniform sampling of the analog front-end output signal (which already includes the rectification offset component). Let the ideal sampling interval be... The actual sampling interval is ,in For the first The timing jitter of each sampling period results in the sampling value error. An approximate formula is: ;

[0119] When input signal Includes low-frequency components caused by rectification offset Although the slope of this low-frequency component is relatively small, the sampling error it causes in high-precision control systems (such as ADCs with 16-bit or higher resolution) is still not negligible. This is especially true for rectified offset. Timing jitter when exhibiting slow drift characteristics The product of the signal slope and the noise level will produce wideband sampling noise, the power of which is related to the jitter variance. It is proportional to the square of the signal slope.

[0120] In some circuit topologies, the rectification-jitter positive feedback loop forms a positive feedback loop between the two layers mentioned above. This positive feedback loop can cause a sharp degradation or even instability in the control system performance when the radio frequency interference intensity exceeds a certain critical threshold. The core function of the cross-domain coupling model is to capture the gain of this positive feedback loop, thereby intervening in advance before the loop gain approaches a critical value.

[0121] Construct a cross-domain coupling model, build a multidimensional probability space and calculate the joint probability density function, using the nonlinear rectified offset. and the timing jitter characteristics (in terms of the amplitude of the radio frequency sensed jitter component) A two-dimensional probability space (represented by ) with input variables. In actual industrial electromagnetic environments, and All are random variables, and their statistical characteristics depend on the time-varying characteristics of radio frequency interference (including the switching duty cycle, frequency drift, amplitude fading, etc. of the interference source).

[0122] The joint probability density function The estimation uses a two-dimensional kernel density estimation method: ,in, The total number of historical observation samples, and The first The observed nonlinear rectification offset and RF induction jitter component amplitude. and These represent the kernel bandwidth in two dimensions. For the kernel function, we choose the Gaussian kernel function: Core bandwidth and The choice is based on Silverman's rule of thumb: ;in, and They are respectively and The sample standard deviation.

[0123] An analytical formula for a cross-domain coupling model is established, and the output of the cross-domain coupling model is characterized by the following formula: Simulated front-end error propagation term This term characterizes the error contribution generated at the output of the control system after the nonlinear rectification offset is transmitted through the signal link of the analog front-end circuit. To simulate the error transfer function of the front-end circuit from the input stage to the output stage, its formula is: ;in, To simulate the DC gain of the front-end circuit, Zero frequency, and The pole frequency is denoted by . This transfer function reflects the amplification or attenuation characteristics of the analog front-end circuit for rectified offsets of different frequency components. When the rectified offset is pure DC, the error propagation gain is . When the rectified offset includes a low-frequency AC component, the gain changes with increasing frequency, and the pattern of change is determined by the zero-pole distribution. Digital link jitter-sensitive transferred term. It characterizes the error contribution generated at the output of the control system after the radio frequency induction jitter component is transmitted through the digital control link. Let be the jitter-sensitive transfer function of the digital control link, and its formula is: ,in, and The numerator and denominator coefficients of the transfer function are... and These are the orders of the numerator and denominator, respectively. For Z-domain complex variables, this transfer function reflects the sensitivity of digital filters, controller algorithms, and digital-to-analog converters in the digital control link to timing jitter. The calibration is performed as follows: Under known jitter injection conditions, the output error response of the digital control link is measured, and the transfer function coefficients are obtained by fitting using a system identification algorithm.

[0124] Cross-domain coupling transitive terms This is the core innovation of this application, which characterizes the contribution of the coupling interaction effect between the nonlinear rectification offset and the radio frequency induction jitter component to the overall system error. The coupling strength coefficient is... It is a cross-domain coupled transfer function that also includes Laplace variables from the simulation domain. and the Z-domain variable of the numeric field This reflects the signal transmission characteristics across the analog-digital hybrid domain.

[0125] Coupling strength coefficient The additional error generated when the unit rectified offset and the unit jitter amplitude act simultaneously is the excess of the sum of the errors when the two act independently. The system's overall error was calibrated using the following experimental method: the integrated error was measured under three conditions: only RF injection to generate rectified offset (where jitter is actively suppressed), only jitter injection (where RF is shielded), and simultaneous RF and jitter injection. Let the measured system overall errors under the three conditions be as follows: , and Then the coupling strength coefficient Calculated by the following formula: ;in, and These are the rectification offset and jitter amplitude during calibration, respectively.

[0126] Cross-domain coupling transfer function The formula is: ;in, For cross-domain coupling gain, and These are the zero and pole frequencies in the analog domain, respectively. and These represent the zero and pole locations in the digital domain. The physical meaning of this mixed-domain transfer function is: the frequency response characteristics of this complete chain, where the rectified offset changes the operating point of the analog circuit → affects the threshold of the digital circuit → generates additional jitter → is transmitted through the digital link → generates output error.

[0127] Integral domain Defined as and The physically feasible region, i.e.: ;in, The maximum rectification offset that the front-end circuit can withstand without permanent damage is determined by the device breakdown voltage and the maximum allowable power dissipation. The maximum jitter amplitude that the digital control link can tolerate while maintaining basic functionality is determined by the timing margin.

[0128] The model parameters of the cross-domain coupling model are updated online by minimizing the deviation between the statistical expected value and the preset control accuracy threshold.

[0129] The objective function for online iterative updates is defined as: ,in, This is the vector of model parameters to be updated. This is the upper limit of the error corresponding to the preset control accuracy threshold. The regularization coefficient is used to limit the step size of parameter updates to prevent model instability. The parameter value is from the previous iteration.

[0130] Parameter updates are performed using constrained gradient descent. ,in, For the first The learning rate for the next iteration. The projection operator projects the updated parameters onto the physically feasible region. Inside (ensure) , (Constraints such as the poles being located within the stability region are satisfied.) Learning rate Adaptive strategy using Adam optimizer: ,in, The initial learning rate, and The momentum decay factor, It is the exponential moving average of the squared gradient. is the numerical stability constant.

[0131] Two specific implementation methods of cross-domain coupling model

[0132] Mechanism-driven implementation, based on Volterra series theory, expands the nonlinear input-output relationship of the analog front-end circuit as follows: ;in, For the first The second-order Volterra kernel function. For the RF rectification effect, the main contribution comes from the second-order kernel. Its corresponding second-order Volterra transfer function in the frequency domain It determines the amplitude and phase of the intermodulation products.

[0133] Introducing timing jitter into the Volterra framework, the generalized input signal is defined as: ,in, The timing jitter is time-varying. Substituting the generalized input signal into the Volterra series expansion yields the output formula, which includes jitter and nonlinear coupling terms. After simplification, the DC and low-frequency components of this formula represent the analytical form of the cross-domain coupling model output, and their coefficients are determined using the aforementioned experimental calibration method.

[0134] A data-driven implementation, based on the known physical mechanisms described above, uses deep neural networks to construct... arrive The mapping relationship is as follows. The network structure is designed as follows: The input layer contains 4 neurons, which respectively receive... , , and The latter two features explicitly introduce nonlinear interaction terms to reflect the physical prior knowledge of cross-domain coupling.

[0135] The hidden layer consists of three fully connected layers, with 32, 16, and 8 neurons in each layer, respectively. The activation function used is the parameterized rectified linear unit (PReLU). ;in, The learnable leakage coefficient is initially set to 0.01.

[0136] The output layer contains one neuron, the activation function is a linear function, and the output is... The predicted value.

[0137] The loss function for network training is the sum of the mean squared error and the physical constraint penalty term: ;in, For training batch size, This is the network prediction value. These are measured values. This represents the physical constraint penalty coefficient. (Penalty term) The network is forced to learn mappings that satisfy the physical constraint that the contribution of coupling terms to the synthesis error is positive; that is, the network must learn... This is to prevent the network from learning mapping relationships that violate physical laws in sparse areas of training data.

[0138] Training data was collected as follows: Radio frequency fields of varying intensities (0.1V / m to 30V / m, in 0.5V / m increments) and frequencies (80MHz to 6GHz, covering common industrial radio frequency bands) were set up in both a radio frequency shielded room and a radio frequency reverberation room, and data were recorded simultaneously. , and system comprehensive error At least 2000 samples should be collected under each condition, and the total training sample size should be no less than [number missing]. .

[0139] Regardless of the implementation method used, the core is to capture the combined effect of nonlinear rectification and timing jitter on the overall error, rather than a simple univariate fitting.

[0140] Generate simulated front-end bias compensation parameters, the simulated front-end bias compensation parameters Used to offset the nonlinear rectification offset Introduced DC and low-frequency error components. The generation of [the value] requires simultaneous consideration of steady-state compensation accuracy and dynamic response speed. The calculation process is as follows:

[0141] Output based on cross-domain coupling model The components contributing to the simulated front-end error are decomposed as follows: ; Calculate the required bias compensation voltage : ,in, Let be the transfer function from the bias compensation signal injection point to the system output. The formula for the closed-loop gain function of the compensation loop is: ;in, , and These are the proportional, integral, and differential gains, respectively. This is the low-pass filter cutoff frequency for the differential term. Integral gain. Ensure zero steady-state error compensation for DC rectification offset, differential gain Improve the dynamic response speed to rapidly changing radio frequency interference.

[0142] Bias compensation voltage The signal is output through a digital-to-analog converter and injected into the bias network of the analog front-end circuit. The injection point is chosen as the bias current source node of the operational amplifier input stage, by superimposing a current source at this node... Compensation currents with opposite polarity and equal amplitude This achieves active cancellation of rectified offset, where This is the equivalent resistance of the bias network.

[0143] Generate digital control timing reconstruction parameters, which are used to adjust sampling times or communication time slots to avoid timing-sensitive windows caused by radio frequency interference. The generation of timing reconstruction parameters includes:

[0144] Positioning of the time-sensitive window. Phase information based on RF-sensed jitter components. This determines the moment when the clock edge offset is at its maximum. The timing-sensitive window is defined as the moment when the clock edge offset exceeds a preset jitter tolerance. Time range: ;in, and Frequency The start and end times of the sensitive window caused by jitter components: ;

[0145] Calculate the optimal sampling time that maximizes the sampling signal-to-noise ratio while avoiding time-sensitive windows. : ;in, For the useful signal at time Power estimation, For broadband noise power, For a moment The jitter equivalent noise power at the location.

[0146] Dynamic reslot rearrangement of communication time slots is implemented for industrial communication interfaces (including RS-485, CAN, and industrial Ethernet). This involves dynamically adjusting the transmission time slots of data frames based on the position of the timing-sensitive window. The rearrangement strategy is as follows: high-priority data frames are sent in the time slots with the least jitter, while low-priority data frames or idle frames are placed within the timing-sensitive window. This reduces the bit error rate caused by radio frequency interference without decreasing communication throughput.

[0147] The simulation front-end bias compensation and digital timing reconstruction are coordinated. The simulation front-end bias compensation parameters and digital control timing reconstruction parameters are not generated independently, but are solved through joint optimization to maximize their combined effect. The objective function for joint optimization is defined as: ;in, and These are the penalty coefficients for bias compensation voltage and timing offset, respectively. This is the nominal sampling time. Penalty item. Limit the amplitude of the compensation voltage to avoid introducing additional power consumption and noise; penalty term. Limit timing offsets to avoid conflicts with other timing-related modules in the system.

[0148] The joint optimization problem is solved using the Sequential Quadratic Programming (SQP) algorithm. The computation time for each optimization iteration does not exceed one control cycle, ensuring that the coordinated control commands can track changes in the radio frequency environment in real time.

[0149] The coordinated control command is sent to the corresponding hardware execution unit, enabling the analog front-end circuit and the digital control link to maintain a preset control accuracy threshold under radio frequency interference. Figure 5 As shown;

[0150] The coordinated control command includes two sets of data: the first set is a bias compensation parameter data package, which includes the compensation voltage value. The first group contains the compensation polarity flag and the digital-to-analog converter range configuration information; the second group is the timing reconstruction parameter data packet, which includes the optimal sampling time offset. Communication slot reordering table and jitter tolerance update value.

[0151] The hardware execution unit includes:

[0152] Analog bias injection circuit: includes a high-precision digital-to-analog converter (resolution not less than 16 bits), a low-pass filter (cutoff frequency 1.5 to 2 times the control bandwidth), and a voltage-to-current conversion stage. The digital-to-analog converter receives the compensation voltage value. It is converted into an analog voltage signal, and after the quantization noise is filtered out by a low-pass filter, the voltage-to-current conversion stage generates a compensation current and injects it into the bias network of the analog front-end circuit.

[0153] Digital timing adjustment circuitry includes a programmable delay line (delay resolution not less than 10 ps, ​​delay range covering at least one full clock cycle), a digitally controlled oscillator (frequency adjustment range covering ±5% of the nominal frequency), and a clock multiplexer. The programmable delay line is adjusted according to the optimal sampling time offset. The phase of the sampling clock is adjusted, and the clock multiplexer switches between different clock sources according to the communication time slot rearrangement table to avoid timing-sensitive windows.

[0154] Coordinated control instructions are transmitted from the control processor to each hardware execution unit via a high-speed serial bus (including SPI, I2C, or a custom LVDS link). The transmission protocol uses a data frame format with cyclic redundancy check (CRC-16) to ensure the integrity of instruction transmission in electromagnetic interference environments. If the CRC check fails, the hardware execution unit retains the parameters of the last valid instruction and reports a communication error to the upper-level controller, which then decides whether to trigger instruction retransmission.

[0155] This embodiment obtains the nonlinear rectification offset and timing jitter characteristics, constructs a cross-domain coupling model based on their joint distribution, and generates coordinated control commands to synchronously complete analog bias compensation and digital timing reconstruction. This can effectively reduce the additional errors caused by analog-digital cross-coupling, stabilize the overall error of the control system under complex time-varying radio frequency interference conditions, suppress sampling mis-triggering, and improve the electromagnetic interference immunity of industrial electronic control systems.

[0156] To verify the radio frequency (RF) signal anti-interference control method for industrial electronic control systems proposed in this application, an industrial electronic control behavioral simulation platform was built using an Intel Xeon-Gold 6330 processor with 128GB of memory and MATLAB-Simulink R2024b. The RF interference waveforms were generated based on the industrial field RF interference characteristic parameters disclosed in the publicly available conference paper (DOI: 10.1109 / ICEMI54254.2021.9623718). A simulation sample set was generated within the simulation platform, covering the 80MHz-6GHz frequency band and including four typical operating conditions: continuous wave interference, amplitude modulation (AM) RF interference, multi-source aliasing RF interference, and pulse RF interference. The simulation model fully simulates the front-end semiconductor nonlinear device model and the clock distribution network RF induction jitter model, which can reproduce physical processes such as RF field coupling, nonlinear rectification offset generation, and clock link timing jitter. Three existing methods were selected as baselines for comparison. Baseline-1 is a well-known traditional decoupled error assessment method in the field of industrial EMI simulation, which calculates the analog domain rectification offset and digital domain timing jitter independently, completely ignoring the cross-coupling effect between analog and digital, and is widely used for electromagnetic interference mechanism simulation of industrial control systems. Baseline-2 is the open-source TSMixer multivariate timing regression network architecture, which directly uses RF field strength and rectification offset timing signals as inputs and outputs a system comprehensive error prediction. This network itself does not introduce physical mechanism constraints. Baseline-3 is the open-source one-dimensional convolutional autoencoder TCAE network architecture, which relies on convolutional feature extraction to complete interference error fitting and is often used for signal reconstruction in RF scenarios. The above three baseline methods were reproduced in this simulation environment and compared with the cross-domain coupling model of this application. 7200 sets of samples were randomly divided from the self-built simulation sample set as training samples and 2400 sets of samples were used as test samples. The RF field strength of the test samples covered the gradient conditions of 5V / m-28V / m. The true value of the system comprehensive error of each set of samples was recorded. The results are directly output from the behavioral-level simulation and used for benchmarking the prediction results of various models. Table 1 below lists all the statistical results obtained from the calculations of this simulation platform.

[0157] Table 1 Comparison of prediction performance of different models on the test set

[0158]

[0159] RMSE (Root Mean Square Error) represents the root mean square error between the predicted value and the true label; the smaller the value, the higher the model's prediction accuracy. MAE (Mean Absolute Error) represents the average absolute error. R² (Coefficient of Determination) represents the model's ability to explain sample variation; the closer it is to 1, the stronger the model's fit. MAPE (Mean Absolute Percentage Error) represents the average percentage of error relative to the true value, reflecting the model's relative prediction bias. STD is the standard deviation of the model's prediction error across all test samples. MAXE is the maximum absolute error per sample in the test set. CORR is the Pearson correlation coefficient between the predicted output and the true value; the closer it is to 1, the better the temporal consistency. INF-TIME is the time taken for single-sample inference computation, in μs, reflecting the model's real-time computational overhead.

[0160] As shown in Table 1, Baseline-1 has the highest RMSE, MAE, MAPE, STD, and MAXE among the four models, and the lowest R² and CORR. Baseline-2 and Baseline-3 show progressively improved performance, but MAXE remains high. Proposed-Ours has the lowest RMSE, MAE, MAPE, STD, and MAXE among the four models, with R² and CORR closest to 1. Its inference time is moderate, balancing prediction accuracy and real-time computing power. All quantitative indicators are superior to the other three comparison models. To further illustrate the differences in the timing of the predicted outputs and the true values ​​of different models, a 1.2-second timing segment of a multi-source aliasing RF interference condition was extracted from the test set. The RF field strength under this condition continuously and dynamically fluctuates between 7V / m and 24V / m. The true value curve and the system comprehensive error prediction timing curves of the four models' outputs under this condition were recorded, as shown below. Figure 6 As shown, the figure contains five time series trajectories: the truth curve, Baseline-1, Baseline-2, Baseline-3, and Proposed-Ours. Figure 6 The time-series trajectory can be observed intuitively. When the radio frequency interference field strength fluctuates rapidly and the coupling effect is significantly enhanced, the traditional decoupling model Baseline-1 will show a significant underestimation phenomenon. The two types of data-driven models, TSMixer and TCAE, have local time-series overfitting or lag phenomena. The cross-domain coupling model of this application outputs a trajectory that closely follows the true time-series trajectory throughout the entire process, and its dynamic following ability is significantly better than the comparison baseline.

[0161] Relying solely on the overall test set metrics and timing curves is insufficient to observe the performance differentiation patterns of each model under different RF field strength levels. Therefore, all 2400 test samples were divided into three field strength ranges: 5-10V / m, 10-18V / m, and 18-28V / m. The RMSE (Recovery Mean Squared Error) of each model within each field strength range was statistically analyzed. The RF field strength range was plotted on the x-axis, and the average RMSE of the range was plotted on the y-axis. Figure 7 As shown in Figure 2, four sets of bars are arranged side by side for each field strength interval, corresponding to the interval root mean square error statistics of the four types of models. From the bar distribution in Figure 2, it can be seen that in the low field strength condition of 5-10V / m, the error difference between the various models is relatively small, and the analog-digital cross-coupling effect is relatively weak in this condition. As the RF field strength increases to the medium field strength interval of 10-18V / m, the cross-coupling error gradually becomes prominent, and the error of the traditional decoupling model Baseline-1 rises rapidly. When entering the high RF field strength interval of 18-28V / m, the rectification-jitter positive feedback effect begins to appear, and the RMSE of Baseline-1 increases sharply to 26.41mV. TSMixer and TCAE also show obvious performance degradation. However, the cross-domain coupling model of this application benefits from the explicit introduction of the cross-domain coupling transfer term and the two-dimensional joint probability density statistics. Even under high field strength conditions, the RMSE remains at a low level of 6.14mV, which fully demonstrates that the model of this application has a stronger adaptability to strongly coupled RF interference conditions.

[0162] Based on the above model verification, a complete simulation of the entire chain of radio frequency signal anti-interference control for industrial electronic control systems was run. This included the entire process from orthogonal dual-frequency radio frequency detection signal injection, nonlinear rectification offset calculation, timing jitter feature extraction, online iterative update of cross-domain coupling model parameters, generation of collaborative control commands, and hardware unit execution compensation. This verified the improvement in system control accuracy after closed-loop control using the entire method. The simulation interference samples generated by this simulation platform were used, with the simulation model configured with determined analog front-end circuit parameters and clock distribution network parameters. Three different levels of radio frequency interference intensity were set for the simulation: low interference level 8V / The interference levels are: m, medium interference level 16V / m, and high interference level 24V / m. Four operating modes are implemented under each interference level: Mode 1 is the original, non-interference-resistant control mode, without any compensation or timing adjustments; Mode 2 only performs analog front-end bias compensation without digital timing reconstruction; Mode 3 only performs digital timing reconstruction without analog bias compensation; Mode 4 is the complete coordinated control mode of this application, simultaneously performing analog bias compensation and digital timing reconstruction. Each operating condition simulation lasts for 20 seconds, and each condition is repeated 20 times with the average value taken. The closed-loop simulation statistics include RF field strength, operating mode, and steady-state average error. Error standard deviation Maximum instantaneous error The table shows the number of ADC sampling mis-triggers, the average number of iterations in the compensation loop, and the average computational load percentage. All statistical results in Table 2 are from the simulation output of this simulation platform.

[0163] Table 2. System Closed-Loop Performance Statistics under Different Radio Frequency Interference Levels and Control Modes

[0164]

[0165] Analysis of the statistical data in Table 2 shows that enabling either analog bias compensation or digital timing reconstruction alone can reduce the overall system error, lower the maximum instantaneous error, and reduce ADC sampling mis-triggers to a certain extent. However, neither single mode can eliminate the residual error caused by analog-digital cross-coupling. As the RF field strength increases to a high interference level of 24V / m, the average error under single control mode remains above 23mV, the maximum instantaneous error exceeds 48mV, and the number of sampling mis-triggers reaches 35 and 38 times, respectively. In contrast, using the complete coordinated control strategy of Mode 4 in this application, under all interference levels of 8V / m, 16V / m, and 24V / m, the steady-state average error, error standard deviation, maximum instantaneous error, and number of sampling mis-triggers all show significant reductions. Under high field strength conditions, the average number of model iterations increases slightly, but the computational load remains within the acceptable range of the industrial controller. Under strong RF interference of 24V / m, the steady-state average error is controlled to 7.35mV, and the number of ADC sampling mis-triggers within a 20-second simulation cycle is reduced to 6. This demonstrates that only through the synergy of analog domain compensation and time-domain timing reconstruction can the comprehensive error caused by cross-domain coupling be suppressed, highlighting the necessity of the collaborative control command generation logic in this application. To further illustrate the details of the dynamic evolution of the system's comprehensive error over simulation time under different control modes, the most stringent operating condition of 24V / m high RF interference was selected. Timing data for no control, only analog compensation, only digital timing reconstruction, and complete collaborative control were recorded simultaneously. Additionally, timing data for two intermediate transition parameter operating conditions (partial compensation, partial timing adjustment) were superimposed, as shown below. Figure 8 As shown, Figure 8 Include Figure 8 (a) Figure 8 (b) Figure 8 (c); Figure 8 (a) Used to display the error scatter curves of multiple repeated simulations under uncontrolled conditions. Figure 8 (a) Three scatter plots of the system integrated error of the 1st, 8th and 15th repeated simulations were plotted simultaneously. It can be seen that under the condition of no control, multiple simulations showed large random fluctuations, and the instantaneous error peak generally exceeded 60mV. Figure 8 (b) Used to display the timing results for partially compensated operating conditions. Figure 8 (b) Simultaneously plotting curves for only analog bias compensation, only digital timing reconstruction, and 50% analog-digital co-compensation, it can be seen that although various local compensation schemes can reduce some error amplitude, there are still obvious error peaks. Figure 8 (c) Used to demonstrate the complete synergistic regulation and two sets of parameter reduction synergistic control conditions. The three scatter curves of the complete synergistic regulation, coupling coefficient halving synergistic regulation, and transfer gain reduction synergistic regulation in this application. The complete synergistic regulation condition will constrain the overall error to a narrow range of -10mV to +10mV. The errors of the other two sets of parameter reduction control conditions will increase to different degrees.

[0166] After confirming the overall timing error improvement effect, simulation observations were continued on key intermediate physical quantities within this application. In addition to the 24V / m baseline condition, intermediate variable timing data were simultaneously collected under two sets of moderate RF interference conditions at 20V / m and 16V / m to track the nonlinear rectification offset. RF sensing jitter amplitude Coupling strength coefficient Compensation output voltage Loss function As time and number of iterations change, such as Figure 9 As shown, Figure 9 (a) Shows the timing evolution of the nonlinear rectification offset under different RF field strengths, corresponding to three field strengths: 16V / m, 20V / m, and 24V / m. It can be clearly seen that the higher the RF field strength, the greater the drift of the rectified offset; Figure 9 (b) Show the changes in RF jitter amplitude and coupling strength coefficient with external interference. A comparison of RF jitter amplitude, coupling strength coefficient at 24V / m and 20V / m can be obtained. It can be seen that the RF jitter amplitude fluctuates synchronously with the external RF environment, and the coupling strength coefficient drifts with the change of the circuit operating point. There are obvious differences in the baseline level of the coupling coefficient under different field strengths. Figure 9 (c) shows the dynamic waveforms of the analog front-end compensation voltage of the closed-loop output under different interference intensities, with compensation voltages corresponding to 16V / m, 20V / m, and 24V / m. The compensation voltage follows the drift of the rectifier offset to achieve reverse dynamic following compensation; Figure 9(d) shows the convergence characteristics of the loss function during the online iterative update process of the model. The loss function changes with the number of iterations under the three interference conditions of 16V / m, 20V / m, and 24V / m. It can be observed that there are differences in the convergence speed under different radio frequency interference intensities. All conditions are completed within the 35-55 iteration interval. The loss function drops below the convergence threshold, which meets the computation time constraint of real-time closed-loop control and does not exhibit iterative divergence.

[0167] To evaluate the robustness of this method under various operating conditions with different clock frequencies and analog front-end bandwidths, multiple sets of parameter sweep simulations were set up. The clock frequencies ranged from 25MHz, 50MHz, 100MHz, and 150MHz, and the analog front-end signal bandwidth ranged from 2kHz, 10kHz, 50kHz, and 200kHz. The radio frequency interference field strength was fixed at 18V / m, representing a moderate to strong interference level. The complete collaborative control scheme of this application was used in all operating conditions, and the steady-state average error of the system under different parameter combinations was statistically analyzed. ,like Figure 10 As shown, the color intensity of each block corresponds to the magnitude of the steady-state average error under the corresponding set of parameters. The specific error value for the corresponding operating condition is attached next to the block. It can be seen that when the clock frequency increases and the analog front-end bandwidth increases, the overall system error will increase slightly. However, the steady-state average error remains within 8.6mV under all scanning conditions, without any precipitous performance degradation. This indicates that the entire method of this application has good robustness to common hardware parameter changes in industrial electronic control systems. The algorithm does not have the limitation of only adapting to extremely narrow hardware parameter ranges and has the basic conditions for engineering implementation.

[0168] This embodiment shows that the prediction accuracy of the cross-domain coupling model is better than that of the existing decoupling mechanism method and the two types of data-driven baseline models. The complete collaborative control process can simultaneously suppress the nonlinear rectification offset in the analog domain and the radio frequency induction jitter in the digital domain, effectively reducing the additional error caused by analog-digital cross-domain coupling.

[0169] Example 3 describes in detail a radio frequency signal interference suppression and control system for industrial electronic control systems, such as... Figure 11 As shown, it includes a multi-source sensing module, a cross-domain coupling modeling module, a collaborative control and decision-making module, and a hardware execution module, specifically:

[0170] The multi-source sensing module acquires the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of radio frequency field, as well as the timing jitter characteristics of the digital control link in the current electromagnetic environment.

[0171] The cross-domain coupling modeling module constructs a cross-domain coupling model that characterizes the contribution of radio frequency interference to the overall error of the control system based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics.

[0172] The collaborative control decision module generates collaborative control instructions, including analog front-end bias compensation parameters and digital control timing reconstruction parameters, based on the output of the cross-domain coupling model.

[0173] The hardware execution module sends the coordinated control command to the corresponding hardware execution unit, so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

[0174] The multi-source sensing module, serving as the front-end input unit of the entire system, is responsible for the acquisition, preprocessing, and calculation of two types of physical quantities: nonlinear rectification offset and timing jitter characteristics. Internally, the module includes a radio frequency (RF) probe signal generation subunit, a high-speed analog-to-digital converter (A / D) subunit, a multi-channel time-to-digital converter (TDC) subunit, and a signal preprocessing subunit. The RF probe signal generation subunit outputs orthogonal dual-frequency RF probe signals. The detection signal is injected into the input stage of the analog front-end circuit; the high-speed analog-to-digital converter subunit acquires the baseband response signal at the output of the analog front-end. The signal preprocessing subunit performs band-stop filtering and coherent demodulation operations to obtain the amplitude of the second-order intermodulation products. Amplitude of third-order intermodulation products Combined with the injected power Load impedance Second-order nonlinear coefficients Third-order nonlinear coefficients Environmental Coupling Correction Factor Complete the calculation and output the nonlinear rectified offset. Meanwhile, the multi-channel time-to-digital converter (TDC) subunit reads data from multiple jitter sampling nodes on the clock distribution network and outputs an edge offset sequence. The signal preprocessing subunit performs detrending and Kaiser windowing on the edge offset sequence. Windowing, discrete Fourier transform, power spectrum estimation calculations, combined with the number of sampling points RF field power spectral density Background noise power spectral density and regularization constant Determine the power spectral density of the radio frequency inductive jitter component. Simultaneously, the phase information of the radio frequency induction jitter component is calculated. The multi-source sensing module outputs preprocessed structured samples, with each set of samples carrying the corresponding time-to-time data. , , The sample sequences are sent to the cross-domain coupled modeling module for model building and parameter updating, and are also stored on the local non-volatile storage medium to accumulate historical observation samples for subsequent kernel density estimation operations.

[0175] The cross-domain coupled modeling module receives the output from the multi-source sensing module. Timing jitter characteristics This module internally includes a probability and statistics subunit, a cross-domain error analysis subunit, and a parameter iterative optimization subunit. The probability and statistics subunit reads the historical observation sample set and uses the total number of samples... Two-dimensional kernel bandwidth , and Gaussian kernel function Perform a two-dimensional kernel density estimation operation to obtain the two-dimensional joint probability density function. Cross-domain error analysis subunit loads analog domain error transfer function. Digital domain jitter-sensitive transfer function Cross-domain coupling transfer function Combined with the integration domain For the instantaneous comprehensive error function With joint probability density Perform integral calculations to determine the statistical expectation value of the overall error of the control system. Instantaneous synthesis error function Depend on , Coupling strength coefficient Together with three types of transfer functions, it constitutes the parameter iterative optimization subunit with the target loss function. Minimize the expected value as the optimization objective; the loss function is derived from the statistical expectation value. Preset error limit Regularization coefficient and the parameter vector from the previous round Composition; Execution of the projection operator with feasible region Constrained Adam gradient descent iterations, with the learning rate used in the iteration process. gradient vector After the iteration converges, update the vector of all undetermined parameters of the model. parameter vector Includes coupling strength coefficient Cross-domain coupling gain Simulation domain zero and pole parameters , Digital domain zero and pole parameters , The cross-domain coupling modeling module offers two optional implementation paths. The first is the mechanism-driven path, based on the Volterra kernel function. The first approach involves analyzing the coupling relationship between nonlinearity and jitter; the second approach is a data-driven path, using a multi-layer fully connected neural network with PReLU activation units to handle the input. To output The mapping operation uses one of two implementation paths, and the module ultimately outputs the statistical expected value. The identified internal parameters of the model are then processed, and the results are transmitted to the collaborative control and decision-making module.

[0176] The collaborative regulation decision-making module receives the output from the cross-domain coupling modeling module. The model coupling parameters are simultaneously received from the phase information of the multi-source sensing module. The module internally includes simulation compensation solution sub-units, time-series reconstruction solution sub-units, and joint optimization solution sub-units. The simulation compensation solution sub-unit is based on... Decomposition to obtain the simulated front-end error contribution components Combined with the compensation point transfer function Compensation loop closed-loop gain function Solving for the simulated front-end bias compensation parameters The timing reconstruction solution sub-unit is based on phase information. jitter tolerance Determine the timing-sensitive window Under the constraint of avoiding the timing-sensitive window, combined with signal power Noise power jitter equivalent noise power Solving for the optimal sampling time Further calculate the optimal sampling time offset , The optimal sampling time Subtract the nominal sampling time The result is obtained. The joint optimization solution sub-unit utilizes the Sequential Quadratic Programming (SQP) algorithm to achieve the desired result. , As a penalty coefficient, for and Joint constraint optimization is performed, and the encapsulated coordinated control instruction data packet is output. The data packet is divided into bias compensation parameter data packet and timing reconstruction parameter data packet, which respectively carry... Compensation polarity flag, DAC range configuration information, and The communication time slot reordering table and jitter tolerance update value are used to send the complete data packet to the hardware execution module.

[0177] The hardware execution module receives the coordinated control instruction data packets output by the coordinated control decision module. The module includes a bus communication subunit, an analog bias injection subunit, and a digital timing adjustment subunit. The bus communication subunit parses the data packets via SPI, I2C, or a custom LVDS high-speed serial bus and performs CRC-16 cyclic redundancy check. If the check passes, it will... The signal is delivered to the analog bias injection subunit. The communication time slot reordering table is sent to the digital timing adjustment subunit; if the CRC check fails, the previous version's valid parameters are maintained unchanged and the fault status is reported to the upper layer. The analog bias injection subunit uses a high-precision digital-to-analog converter to... Converted to analog level, then low-pass filtered and voltage-to-current converted to generate a compensation current. Compensation current Depend on Equivalent resistance of bias network The calculated bias current source node of the compensated current injection into the analog front-end circuit is then determined. The digital timing adjustment subunit reads the optimal sampling time offset. It drives the programmable delay line to complete the clock phase shift, and controls the clock multiplexer according to the communication time slot reorder table to complete the communication time slot switching, so that the analog front-end circuit and digital control link maintain the preset control accuracy threshold in the current radio frequency electromagnetic environment.

[0178] This application also provides electronic devices suitable for use in this application, such as... Figure 12 As shown, the electronic device includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores computer program instructions that can be executed by the processor. When the program instructions are executed, the processor performs all the software logic operations to implement the aforementioned multi-source sensing module, cross-domain coupling modeling module, and collaborative control decision module, and calls the hardware execution module to complete the hardware-level control actions.

[0179] This application also provides a corresponding computer-readable storage medium, on which computer instructions are stored. When the computer instructions are executed by a processor, they perform all the steps of the radio frequency signal anti-interference control method for industrial electronic control systems described in this application. The storage medium internally stores parameter storage space corresponding to each variable character involved in the calculation, including space for storing... Variable cache and storage timing jitter characteristics Cache area, storage model parameter vector Parameter storage area, storage and The output parameter buffer is used to read and modify variables in the buffer during program instruction execution to complete all logical operation processes.

[0180] This embodiment, based on multi-source sensing, cross-domain coupling modeling, collaborative control decision-making, and the coordinated work of hardware execution modules, can collect two types of characteristic quantities induced by radio frequency interference in real time, identify cross-domain coupling parameters online and output appropriate control commands, and the hardware unit synchronously performs compensation and timing adjustment, robustly adapting to different hardware parameter conditions, and reliably realizing radio frequency interference immunity control of industrial electronic control systems.

[0181] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any changes, modifications, substitutions, integrations, and parameter changes made to these embodiments within the spirit and principles of the present invention, without departing from the principles and spirit of the present invention, through conventional substitutions or to achieve the same function, fall within the scope of protection of the present invention.

Claims

1. A method for suppressing interference and controlling radio frequency signals in industrial electronic control systems, characterized in that, include: To obtain the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of radio frequency field, and the timing jitter characteristics of the digital control link under the current electromagnetic environment; Based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics, a cross-domain coupling model is constructed to characterize the contribution of radio frequency interference to the overall error of the control system. Based on the output of the cross-domain coupling model, a coordinated control command is generated, which includes analog front-end bias compensation parameters and digital control timing reconstruction parameters. The analog front-end bias compensation parameters are used to offset the DC and low-frequency error components introduced by the nonlinear rectification offset, and the digital control timing reconstruction parameters are used to adjust the sampling time or communication time slot to avoid the timing-sensitive window caused by radio frequency interference. The coordinated control command is sent to the corresponding hardware execution unit so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

2. The method for anti-interference control of radio frequency signals in industrial electronic control systems according to claim 1, characterized in that, The acquisition of the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of a radio frequency field includes: An orthogonal dual-frequency radio frequency detection signal is injected into the input stage of the analog front-end circuit, the frequency of which is located in the out-of-band frequency band of the industrial electronic control system; The baseband response signal at the output of the analog front-end circuit is acquired, and the intermodulation components of the baseband response signal are extracted to obtain the amplitude of the second-order intermodulation product and the amplitude of the third-order intermodulation product. The nonlinear rectification offset is calculated based on the amplitude of the second-order intermodulation product, the amplitude of the third-order intermodulation product, and the injection power of the orthogonal dual-frequency radio frequency detection signal.

3. The method for anti-interference control of radio frequency signals in industrial electronic control systems according to claim 2, characterized in that, The nonlinear rectification offset is calculated based on the amplitudes of the second-order intermodulation products, the third-order intermodulation products, and the injected power of the orthogonal dual-frequency radio frequency detection signal, using the following formula: ;in, This refers to the nonlinear rectification offset. The injection power of the orthogonal dual-frequency radio frequency detection signal. For load impedance, and These are the second-order and third-order nonlinear coefficients of the analog front-end circuit, respectively. and These are the amplitudes of the second-order intermodulation product and the amplitudes of the third-order intermodulation product, respectively. The frequency of the orthogonal dual-frequency radio frequency detection signal , Related environmental coupling correction factors.

4. The method for anti-interference control of radio frequency signals in industrial electronic control systems according to claim 1, characterized in that, The acquisition of timing jitter characteristics of the digital control link under the current electromagnetic environment includes: Multiple jitter sampling nodes are set in the clock distribution network of the digital control link, and the edge offset sequence of each jitter sampling node is collected in multiple consecutive clock cycles. The edge offset sequence is decomposed in the frequency domain to extract periodic jitter components and random jitter components; Based on the spectral peak distribution of the periodic jitter components, the radio frequency induced jitter components related to the frequency of the interference source in the radio frequency field are identified, and the amplitude and phase information of the radio frequency induced jitter components are used as the timing jitter features.

5. The method for anti-interference control of radio frequency signals in industrial electronic control systems according to claim 4, characterized in that, Based on the spectral peak distribution of the periodic jitter components, the radio frequency induced jitter components related to the frequency of the interference source in the radio frequency field are identified, as shown in the formula: in, For frequency The power spectral density of the radio frequency sensed jitter component at that location. The first edge offset sequence Each sample value, The number of sampling points. For window functions, The source of interference in the radio frequency field at frequency Power spectral density estimation at [location] The background noise power spectral density of the digital control link is given. is the regularization constant.

6. The method for anti-interference control of radio frequency signals for industrial electronic control systems according to claim 1, characterized in that, The cross-domain coupling model, which characterizes the contribution of radio frequency interference to the overall error of the control system, is constructed based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics, including: A multidimensional probability space is constructed with the nonlinear rectification offset and the timing jitter characteristics as input variables, and the joint probability density function of the two is calculated in the multidimensional probability space. Based on the joint probability density function, the cross-domain coupling model is established, and the output of the cross-domain coupling model is the statistical expectation value of the comprehensive error of the control system. The model parameters of the cross-domain coupling model are updated online by minimizing the deviation between the statistical expected value and the preset control accuracy threshold.

7. The method for anti-interference control of radio frequency signals in industrial electronic control systems according to claim 6, characterized in that, The output of the cross-domain coupling model is the statistical expectation value of the integrated error of the control system, and the formula for the statistical expectation value is: in, This is the statistical expectation value of the overall error of the control system; The instantaneous comprehensive error function is determined by the analog front-end error propagation relationship, the digital link jitter propagation relationship, and the analog-digital cross-coupling relationship. It is the joint probability density function of nonlinear rectified offset and RF induction jitter component; This is the integration domain.

8. A radio frequency signal anti-interference control system for industrial electronic control systems, applicable to the radio frequency signal anti-interference control method for industrial electronic control systems as described in any one of claims 1-7, characterized in that, It includes a multi-source sensing module, a cross-domain coupling modeling module, a collaborative control and decision-making module, and a hardware execution module, specifically: The multi-source sensing module acquires the nonlinear rectification offset of the analog front-end circuit in the industrial electronic control system under the action of the radio frequency field, as well as the timing jitter characteristics of the digital control link in the current electromagnetic environment. The cross-domain coupling modeling module constructs a cross-domain coupling model that characterizes the contribution of radio frequency interference to the overall error of the control system based on the joint distribution of the nonlinear rectification offset and the timing jitter characteristics. The collaborative control decision module generates a collaborative control instruction that includes analog front-end bias compensation parameters and digital control timing reconstruction parameters based on the output of the cross-domain coupling model. The hardware execution module sends the coordinated control command to the corresponding hardware execution unit, so that the analog front-end circuit and the digital control link maintain a preset control accuracy threshold under radio frequency interference environment.

9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the radio frequency signal anti-interference control method for industrial electronic control systems as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to execute the radio frequency signal anti-interference control method for industrial electronic control systems as described in any one of claims 1 to 7.

Images