A large inertia space two-dimensional turntable resonance suppression system and method based on Goertzel notch filter

Abstract

The present application relates to low-speed large torque motion control technical field, especially a kind of big inertia space two-dimensional rotary table resonance suppression system and method based on Goertzel notch filter.The system of the present application includes servo three-ring control, rotational speed acquisition, resonance identification, notch filter and parameter updating unit, and forms position ring, speed ring and current ring nested three-ring control architecture.The method of the present application is applied to the system, resonance identification and suppression strategy are embedded in speed ring control channel, through the acquisition speed ring real-time feedback rotational speed time domain signal, through the improved Goertzel algorithm combined with adaptive threshold golden section search, complete the online identification of mechanical resonance characteristic parameters, output real-time resonance characteristic parameters.Meanwhile, based on resonance characteristic parameters, pre-distortion bilinear transformation method is used to design dynamic adjustable notch filter, and the frequency offset error of filter discretization is eliminated.The present application can accurately suppress mechanical resonance under large load working condition, significantly reduce resonance peak value, and improve system control bandwidth simultaneously.

Description

Technical Field

[0001] This invention relates to the field of low-speed, high-torque motion control technology, and in particular to a large-inertia spatial two-dimensional turntable resonance suppression system and method based on a Goertzel notch filter. Background Technology

[0002] In space-to-ground or inter-satellite optical communication, two-dimensional turntables, characterized by their large rotation range and simple control, are a crucial component of the pointing mechanism in spaceborne ATP systems. The control performance of the turntable directly affects the dynamic response, tracking accuracy, and operational stability of the entire system. Due to structural constraints, the connection stiffness between the two-dimensional turntable and the optical telescope structure is often low, resulting in a high inertia ratio electromechanical coupling characteristic in the drive system. Mechanical resonance and limited control bandwidth are two particularly prominent and interconnected challenges. At specific frequencies, these can cause strong oscillations in the output response, restricting the system's tracking capability and its ability to suppress external disturbances.

[0003] Common control methods reduce the system gain of the velocity loop by changing controller parameters or controller structure. These methods include sliding mode control (SMC), model predictive control (MPC), active disturbance rejection control (ADRC), accelerometer, and state feedback-based proportional-integral control (SFPIC). These control methods are often effective in suppressing jitter in situations with small loads, but they are somewhat insufficient for large-inertia space telescopes. Summary of the Invention

[0004] The purpose of this invention is to provide a large-inertia spatial two-dimensional turntable resonance suppression system and method based on Goertzel notch filter, which mainly solves the problems existing in the prior art, namely, the mechanical resonance caused by motor jitter under large inertia and the limited improvement of control bandwidth, and provides an efficient and widely applicable solution for coarse tracking systems in space optical communication.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is to provide a resonance suppression control system for a large inertia spatial two-dimensional turntable, characterized in that it includes a servo three-ring control unit, a rotation speed acquisition unit, a resonance identification unit, a notch filter unit, and a parameter update unit.

[0006] The servo three-loop control unit implements a servo three-loop control architecture, consisting of a position loop, a speed loop, and a current loop from the outside in. The speed acquisition unit acquires the speed of the speed loop and provides real-time feedback of the speed time-domain signal. The resonance identification unit is connected to the speed acquisition unit, acquires the speed time-domain signal, and uses an improved Goertzel algorithm combined with an adaptive threshold golden section search strategy to identify and output resonance characteristic parameters online. The notch filter unit is a dynamically adjustable notch filter connected to the resonance identification unit. Based on the resonance characteristic parameters, it generates filter parameters using a pre-distortion bilinear transform method to filter the torque command output by the speed loop. The parameter update unit is connected to the resonance identification unit and uses a sliding window Goertzel algorithm to monitor the changes in the resonance characteristic parameters in real time and dynamically update the core filter parameters of the notch filter unit online.

[0007] Furthermore, the sampling frequency of the current loop is higher than that of the speed loop; the sampling frequency of the speed loop is higher than that of the position loop; the speed acquisition unit, the resonance identification unit, the notch filter unit, and the parameter update unit operate synchronously with the speed loop and do not interfere with the positioning logic of the position loop and the response characteristics of the current loop.

[0008] Furthermore, the control system is applied to the ATP servo control system of a spaceborne large-inertia two-dimensional photoelectric turntable; the servo three-ring control unit, the rotation speed acquisition unit, the resonance identification unit, the notch filter unit, and the parameter update unit are integrated into a programmable digital logic device and implemented in parallel using hardware logic circuits; the system is suitable for dual-axis servo control and resonance suppression of the positioning and pitch axes.

[0009] This invention also discloses a resonance suppression method for a large inertia two-dimensional spatial turntable based on a Goertzel notch filter, applied to the resonance suppression control system of the aforementioned large inertia two-dimensional spatial turntable, characterized by comprising the following steps:

[0010] Step S100: Acquire the time-domain signal of the rotational speed fed back in real time by the rotational speed acquisition unit;

[0011] Step S200: By improving the Goertzel algorithm and combining it with an adaptive threshold golden section search strategy, the rotational speed time-domain signal is identified online, and the resonance characteristic parameters are output.

[0012] Step S300: Based on the resonant characteristic parameters, the core filter parameters are dynamically generated using the pre-distortion bilinear transform method to construct the notch filter;

[0013] Step S400: The notch filter is embedded in series into the control path of the speed loop to filter the torque command output by the speed loop;

[0014] In step S500, the Goertzel sliding window algorithm is used to monitor the changing trend of the resonant characteristic parameters in real time; when the changing trend exceeds a preset threshold, the process returns to step S200 to re-identify online in order to update the core filter parameters.

[0015] Furthermore, in step S200, the improved Goertzel algorithm employs a second-order IIR recursive structure; the second-order IIR recursive structure calculates the amplitude and phase only for the target resonant frequency point in the rotational speed time-domain signal.

[0016] Further, in step S200, the adaptive threshold golden section search strategy includes:

[0017] Step S201: Use second-order polynomial fitting to remove the trend term in the rotational speed time domain signal, and estimate the noise intensity by median absolute deviation to establish a noise benchmark;

[0018] Step S202: Based on the noise reference and the global measured signal-to-noise ratio of the rotational speed time domain signal, dynamically and adaptively update the threshold coefficient to set a judgment threshold for distinguishing between reliable signals and noise.

[0019] Step S203: Iteratively search for the resonance interval based on the judgment threshold and the golden ratio, and use two levels of sampling points, coarse sampling or fine sampling, depending on the current interval type; use coarse sampling to speed up in the noise interval, and use fine sampling to maintain accuracy in the resonance interval.

[0020] Step S204: Using a weak decision counting and multi-window collaboration mechanism, multi-window joint determination is performed based on the results of the coarse sampling and the fine sampling to achieve global optimal identification of the resonant frequency and output the resonant characteristic parameters.

[0021] Furthermore, in step S300, the pre-distortion bilinear transform method first corrects the resonant frequency shift through pre-distortion compensation, and then completes the filter discretization through bilinear transform, so as to ensure that the core filtering parameters of the discretized notch filter are consistent with those in the continuous domain.

[0022] Furthermore, in steps S200 and S300, the resonant characteristic parameters include the resonant center frequency, resonant amplitude, and resonant bandwidth; the core filtering parameters of the notch filter include the notch center frequency, notch depth, and notch width.

[0023] Further, in step S500, the sliding window Goertzel algorithm includes:

[0024] Step S501: Set a sliding window of fixed length to buffer the rotational speed time-domain signal collected in multiple consecutive control cycles;

[0025] Step S502: Update the rotational speed time domain signal in the sliding window in real time, and remove the historically outdated rotational speed time domain signal;

[0026] Step S503: Based on the historical rotational speed time-domain signal sequence in the sliding window, the Goertzel algorithm is used to calculate the resonance characteristic parameters at each historical moment and generate the resonance parameter change trend.

[0027] Step S504: Based on the trend of the resonant parameters, if the trend exceeds the preset threshold, proceed to step S200 to re-identify online; otherwise, proceed to step S501 to continue monitoring.

[0028] Furthermore, in step S500, the preset threshold is dynamically set based on the global measurement signal-to-noise ratio of the rotational speed time-domain signal.

[0029] In view of the above technical features, the present invention is a turntable control system and method for integrating resonant frequency identification and suppression of Goertzel-notch filter. Compared with the prior art, the present invention has the following significant advantages: the present invention can quickly and accurately identify the center frequency, width and depth of system resonance automatically; and improve the dynamic performance of the two-dimensional turntable control system by obtaining the system error state and adjusting parameters online.

[0030] To verify the effectiveness of the technical solution of this invention, a comparative experiment was conducted on the ATP servo control system of a spaceborne large-inertia two-dimensional photoelectric turntable. In the experiment, the original control method (without the resonance suppression module of this invention) and the method of this invention (with the Goertzel notch filter resonance suppression module) were compared, and the velocity response curves of the azimuth and pitch axes were recorded. The experimental results are as follows: Figure 5 As shown.

[0031] Depend on Figure 5 It is known that when using the original method, there are obvious periodic resonant oscillations in the velocity responses of the azimuth and pitch axes, with large velocity fluctuation amplitudes and long convergence times, which seriously affect the control accuracy and stability margin of the turntable. After adopting the method of the present invention, the velocity response curves of the azimuth and pitch axes are smooth, the resonant oscillations are quickly and accurately suppressed, the velocity response converges rapidly to a steady state, and the steady-state accuracy is significantly better than that of the original method. This effectively verifies the resonance suppression effect and control bandwidth improvement capability of the present invention under large inertia and weak damping conditions. Attached Figure Description

[0032] Figure 1This is a schematic diagram of a preferred embodiment of the large inertia spatial two-dimensional turntable resonance suppression system based on the Goertzel notch filter of the present invention.

[0033] Figure 2 This is a flowchart of a preferred embodiment of the method for suppressing large inertia spatial two-dimensional turntable resonance based on Goertzel notch filter according to the present invention.

[0034] Figure 3 yes Figure 2 Flowchart of the adaptive threshold golden section search method in step S200;

[0035] Figure 4 This is a schematic diagram of the data flow and module interaction of the large inertia spatial two-dimensional turntable resonance suppression method based on Goertzel notch filter according to the present invention;

[0036] Figure 5 This is a schematic diagram illustrating the effect of a preferred embodiment of the method for suppressing large inertia spatial two-dimensional turntable resonance based on Goertzel notch filter according to the present invention.

[0037] In the diagram: 100 - Servo three-ring control unit; 300 - Turntable;

[0038] 110 - Position loop, 120 - Speed ​​loop, 130 - Current loop;

[0039] 210 - Rotation speed acquisition unit, 220 - Resonance identification unit, 230 - Notch filter unit, 240 - Parameter update unit. Detailed Implementation

[0040] The present invention will be further described below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0041] Example 1: A large inertia spatial two-dimensional turntable resonance suppression control system based on Goertzel notch filter.

[0042] Please see Figure 1 This invention discloses a large-inertia spatial two-dimensional turntable resonance suppression control system based on a Goertzel notch filter. A preferred embodiment of this system, used to implement the resonance suppression method described in Embodiment 2 below, includes a servo three-ring control unit 100, a speed acquisition unit 210, a resonance identification unit 220, a notch filter unit 230, and a parameter update unit 240.

[0043] The servo three-loop control unit 100 implements a servo three-loop control architecture, consisting of a position loop 110, a speed loop 120, and a current loop 130, from the outside in. Each loop has different control logic and function. Specifically, the position loop 110 takes the deviation between the commanded position and the actual feedback position as input, calculates the given speed through a position regulator, and outputs a given speed to achieve precise turntable positioning and trajectory tracking. The speed loop 120 receives the speed setpoint output from the position loop 110, compares it with the actual speed feedback to form a speed deviation, and outputs a torque command after calculation by the speed regulator. The current loop 130 receives the current setpoint corresponding to the torque command output from the speed loop 120, quickly adjusts the motor phase current, suppresses current disturbances, ensures stable torque output, and provides underlying power support for the speed and position loops. In the specific implementation, any suitable tuning method can be adopted according to the actual motor parameters and control requirements.

[0044] The speed acquisition unit 210 acquires the speed of the speed loop 120 and provides real-time feedback of the speed time-domain signal. In the three-loop control system, the feedback speed signal from the speed loop 120 directly reflects the disturbance information of mechanical resonance. The resonance identification unit 220 is connected to the speed acquisition unit 210, acquires the speed time-domain signal, and uses an improved Goertzel algorithm combined with an adaptive threshold golden section search strategy to identify and output resonance characteristic parameters online. The resonance characteristic parameters include the resonance center frequency, resonance amplitude, and resonance bandwidth. The notch filter unit 230 is a dynamically adjustable notch filter connected to the resonance identification unit 220. Based on the resonance characteristic parameters, it uses a pre-distortion bilinear transform method to generate filter parameters and filters the torque command output by the speed loop 120. The core filter parameters of the notch filter include the notch center frequency, notch depth, and notch width. The parameter update unit 240 is connected to the resonance identification unit 220 and uses a sliding window Goertzel algorithm to monitor the changes in the resonance characteristic parameters in real time and dynamically update the core filter parameters of the notch filter unit 230 online. The window length of a sliding window is usually determined by a combination of the rate of change of the system's resonant parameters and the processor's computing power. For example, when the changes are rapid, the window needs to be shortened to improve the response speed, while when the changes are slow, the window can be lengthened to improve the frequency resolution.

[0045] Please see Figure 4 The data flow and module interaction of the resonance suppression method of this invention are as follows: The speed reference input is processed by the speed controller (speed loop 120) to form a speed command. The torque command output by the speed loop 120 is used to identify the resonance characteristic parameters, namely the resonance center frequency, resonance amplitude, and resonance bandwidth, by the sliding Goertzel algorithm. The optimal notch parameters are determined by the golden section search strategy. After parameter tuning, the parameters are sent to the notch filter (notch filter unit 230) to filter the torque command, and finally drive the current loop 130 and the turntable 140 to form a closed-loop control.

[0046] The sampling frequency of the current loop 130 is higher than that of the speed loop 120, which in turn is higher than that of the position loop 110. In other words, in the servo three-loop control architecture, the current loop 130 has the highest sampling frequency, followed by the speed loop 120, while the position loop 110 has the lowest. Simultaneously, the speed acquisition unit 210, resonance identification unit 220, notch filter unit 230, and parameter update unit 240 all operate synchronously with the speed loop 120, without interfering with the positioning logic of the position loop 110 or the response characteristics of the current loop 130.

[0047] The control system in this embodiment is applied to the ATP servo control system of a spaceborne large-inertia two-dimensional photoelectric turntable. Its servo three-loop control unit 100, speed acquisition unit 210, resonance identification unit 220, notch filter unit 230, and parameter update unit 240 are all integrated within a programmable digital logic device and implemented in parallel using hardware logic circuits. This control system is suitable for dual-axis servo control and resonance suppression of the positioning and pitch axes, and is used to drive the spaceborne photoelectric turntable to achieve rapid target acquisition and precise pointing.

[0048] In one specific application, the programmable digital logic device is implemented using an FPGA chip. The servo inverter switching frequency is 10kHz, and the position loop feedback frequency is 40Hz. The FPGA chip acquires the turntable position feedback value by collecting the angle signal of the rotary transformer at a sampling frequency of 2kHz. Internally, the FPGA uses the deviation between the given reference position and the position feedback value as the input for the PID calculation of the position loop 110, and outputs the reference value of the speed loop 120 as the speed loop setpoint. The speed loop 120 compares the actual speed feedback to form a speed deviation, and after PID calculation, outputs a torque command as the setpoint of the current loop 130. The current loop 130 enables fast motor torque response. The speed acquisition unit 210, resonance identification unit 220, notch filter unit 230, and parameter update unit 240 are embedded in series in the control path of the speed loop 120, implementing resonance suppression between the output of the speed loop 120 and the input of the current loop 130. The FPGA chip performs three-loop control operations (position loop 110, velocity loop 120, and current loop 130) and resonance identification and suppression operations in parallel, making it suitable for dual-axis servo control and resonance suppression of the position and pitch axes.

[0049] Example 2: A method for suppressing large inertia spatial two-dimensional turntable resonance based on Goertzel notch filter.

[0050] Please see Figure 2This invention also discloses a method for suppressing resonance in a large-inertia two-dimensional spatial turntable based on a Goertzel notch filter. A preferred embodiment is applied to the resonance suppression control system of a large-inertia two-dimensional spatial turntable as described in Embodiment 1. This embodiment uses a three-loop control architecture as an example for illustration. The establishment of the three-loop control architecture includes the steps of establishing a single-axis dual-inertia elastic dynamic model of the large-inertia two-dimensional spatial turntable, and the steps of constructing a servo three-loop control architecture.

[0051] The differential equations of motion for the dual-inertia elastic dynamics model are as follows:

[0052]

[0053] Among them, J M J L θ represents the moment of inertia of the motor rotor and the load turntable, respectively. M θ L These are the rotation angles at the motor end and the load end, respectively; ω M ω L These are the angular velocities at the motor end and the load end, respectively; B M K is the viscous damping coefficient at the motor end. s T represents the torsional stiffness of the drive shaft. e This refers to the electromagnetic torque of the motor.

[0054] By performing a Laplace transform on the above differential equation, the transfer function of the motor speed relative to the electromagnetic torque can be obtained, and the resonant frequency ω in the undamped state can be derived. ARF and anti-resonant frequency ω NTF :

[0055]

[0056] Define the inertia ratio R=J L / J M Then the resonant frequency can be simplified to:

[0057]

[0058] Under high inertia conditions, with an inertia ratio R > 10, the system exhibits characteristics of low-frequency resonance, weak damping, and high resonance peak value. Resonance disturbances are mainly manifested in velocity response fluctuations. Based on these characteristics, this invention embeds a resonance suppression module into the velocity loop control path, directly suppressing the resonance disturbance at the point where it is most significant.

[0059] Taking the above three-loop control architecture as an example, the resonance suppression method specifically includes the following steps:

[0060] Step S100: Acquire the time-domain signal of rotational speed.

[0061] The speed time-domain signal is acquired in real time from the speed acquisition unit. In the three-loop control architecture, the feedback speed signal from the speed loop directly reflects the disturbance information of mechanical resonance. Therefore, this step acquires the real-time feedback speed time-domain signal x[n] from the speed loop as the input for subsequent processing. The acquisition of the speed time-domain signal can be achieved, for example, through a photoelectric encoder in conjunction with an analog-to-digital converter (ADC).

[0062] Step S200: Output the resonant characteristic parameters.

[0063] By improving the Goertzel algorithm and combining it with an adaptive threshold golden section search strategy, the rotational speed time-domain signal is identified online, and the resonance characteristic parameters are output. Figure 4 The data flow and module interaction relationship of this step are shown: the torque command output by the speed loop 120 is processed by the sliding Goertzel algorithm, and the optimal resonance parameters are identified by combining the golden section search strategy.

[0064] The improved Goertzel algorithm employs a second-order IIR recursive structure, calculating amplitude and phase only at the target resonant frequency point in the rotational speed time-domain signal. Compared to the traditional FFT full-band scanning method, this method significantly reduces computational complexity, meeting real-time control requirements. The real-number recursive calculation formula is as follows:

[0065]

[0066] In the above formula, x[n] is the real-time speed feedback time-domain signal acquired by the speed loop, and ω k For the target digital frequency detection, s[n] is a recursive intermediate variable, with the initial condition s[-1]=s[-2]=0. After N-point sampling is completed, the normalized amplitude of the target frequency is calculated through the intermediate variable.

[0067] An adaptive threshold golden section search strategy is used to separate noise frequencies and resonant frequencies. Because the velocity feedback signal contains a mixture of 1 / f low-frequency drift and white noise, directly performing a peak search can easily misinterpret noise as resonance. (See also...) Figure 3 This method employs the following noise-resistant search strategy:

[0068] Step S201: Establish a noise reference.

[0069] A second-order polynomial fitting method is used to remove the trend term in the rotational speed time-domain signal, and the noise intensity is estimated by the median absolute deviation, thereby establishing a noise benchmark and avoiding noise spurious peak interference.

[0070] Step S202: Set the judgment threshold.

[0071] Based on the global measured signal-to-noise ratio (SNR) of the noise baseline and the rotational speed time-domain signal, the threshold coefficient η is dynamically and adaptively updated to set the judgment threshold used to distinguish between reliable signals and noise. The specific value of the threshold coefficient η depends on the SNR level under actual operating conditions, with a typical range of 3 to 10. A higher value is used when the signal is strong to avoid misjudgment, and a lower value is used when the signal is weak to improve detection sensitivity.

[0072] Step S203: Iteratively search for the resonance interval.

[0073] The resonance interval is iteratively searched based on the judgment threshold and the golden ratio of 0.618, and the number of sampling points is adjusted according to the type of the current interval using either coarse sampling or fine sampling. That is, coarse sampling is used to speed up the process in the noisy interval, while fine sampling is used to maintain accuracy in the resonance interval.

[0074] Step S204: Output the resonant characteristic parameters.

[0075] By utilizing a weak decision counting and multi-window collaboration mechanism, a multi-window joint determination is performed based on the results of coarse and fine sampling to achieve globally optimal identification of the resonant frequency. This avoids the problems of incorrect interval discarding and search failure in noisy environments encountered by traditional golden section search algorithms, and finally outputs the resonant characteristic parameters. In this embodiment, the resonant characteristic parameters include the resonant center frequency, resonant amplitude, and resonant bandwidth.

[0076] The weak decision counting mechanism refers to making a final decision only when more than half or a preset proportion of judgments are consistent across multiple independent judgments, thus avoiding misjudgments caused by a single noise spike. The multi-window collaboration mechanism involves simultaneously analyzing multiple sliding windows with different starting points or lengths, combining the judgment results from each window to improve the identification confidence level.

[0077] Step S300: Construct a notch filter.

[0078] Based on the resonant characteristic parameters, the core filter parameters are dynamically generated using the pre-distortion bilinear transform method to construct a notch filter. Specifically, the standard transfer function of the notch filter in the continuous domain is:

[0079]

[0080] Where, ω n The center frequency of the notch filter is matched with the resonant frequency of the velocity path identified in real time by the Goertzel algorithm. k1 is the notch width coefficient, k2 is the notch depth coefficient, and the notch depth Dn = k2 / k1. By constraining the phase lag of the filter through a preset control bandwidth gain threshold, the gain at the servo control bandwidth is guaranteed to be greater than or equal to 0.99, thus balancing the resonance suppression effect with the dynamic performance of the three-loop system.

[0081] To avoid the frequency shift problem caused by conventional bilinear transformation, pre-distortion compensation is used to correct the resonant frequency shift. The specific formula is as follows, where T is the system sampling period:

[0082]

[0083] The filter is then discretized using a bilinear transform, and the discrete-domain notch filter transfer function H is obtained by solving the problem. np (z) ensures that the center frequency and attenuation depth of the discretized filter are consistent with those in the continuous domain, solving the performance degradation problem of high-frequency filters caused by conventional discretization and adapting to the high-frequency real-time control requirements of the speed loop. The specific formula for the bilinear transform is:

[0084]

[0085] Conventional bilinear transforms can produce a frequency warping effect, meaning the actual notch frequency of the discretized filter will be lower than the design value. This invention uses a pre-distortion compensation formula to correct this deviation, ensuring that the discretized notch frequency is precisely consistent with the design value.

[0086] Step S400: Filter the torque command.

[0087] The constructed notch filter is embedded in series in the control path of the speed loop (i.e., between the output of the speed loop regulator and the input of the current loop regulator) to filter the torque command output by the speed loop in real time, thereby accurately canceling the mechanical resonance component of the speed path and outputting the filtered smooth torque command to the current loop.

[0088] Step S500: Monitor the changing trend of the resonant characteristic parameters in real time.

[0089] like Figure 4 As shown, the sliding Goertzel algorithm also serves a "monitoring" function here: after completing the initial identification, the algorithm continues to run, monitoring the changing trend of the resonant characteristic parameters in real time. When the parameter drift exceeds a preset threshold, parameter re-identification is automatically triggered. That is, when the changing trend exceeds the preset threshold, it returns to step S200 to re-perform online identification in order to update the core filter parameters.

[0090] The specific implementation of this step includes the following sub-steps:

[0091] Step S501: Buffer multiple rotational speed time-domain signals.

[0092] A fixed-length sliding window is set to buffer the rotational speed time-domain signal acquired in multiple consecutive control cycles.

[0093] Step S502: Update the rotational speed time domain signal in real time.

[0094] The rotational speed time domain signal in the sliding window is updated in real time, and outdated rotational speed time domain signals are removed.

[0095] Step S503, the trend of resonant parameter changes.

[0096] Based on the historical rotational speed time-domain signal sequence within a sliding window, the Goertzel algorithm is used to calculate the resonance characteristic parameters at each historical moment, generating the trend of resonance parameter changes.

[0097] Step S504: Determine whether to re-perform online identification.

[0098] Based on the trend of the resonant parameters, if the trend exceeds a preset threshold, the process jumps to step S200 to re-identify online; otherwise, it jumps to step S501 to continue monitoring. The preset threshold is dynamically set based on the global measurement signal-to-noise ratio of the rotational speed time-domain signal.

[0099] When the resonant parameter drift is small, the system maintains the current notch filter parameters and continues to operate. When changes in load inertia, damping, and other operating conditions cause a significant drift in the resonant frequency (the trend of change exceeds a preset threshold), the sliding window Goertzel algorithm automatically triggers a return to S200 for re-identification. The notch filter is redesigned with the new resonant parameters, and the center frequency, depth, and width parameters of the notch filter are dynamically corrected to adapt to the time-varying operating conditions of the turntable load damping and inertia. This continuously suppresses speed loop resonance disturbances, thereby achieving full-condition adaptive operation.

[0100] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A resonance suppression control system for a large-inertia spatial two-dimensional turntable, characterized in that, It includes a servo three-ring control unit, a speed acquisition unit, a resonance identification unit, a notch filter unit, and a parameter update unit; The servo three-loop control unit implements a servo three-loop control architecture, consisting of a position loop, a speed loop, and a current loop from the outside in. The speed acquisition unit acquires the speed of the speed loop and provides real-time feedback of the speed time-domain signal. The resonance identification unit is connected to the speed acquisition unit, acquires the speed time-domain signal, and uses an improved Goertzel algorithm combined with an adaptive threshold golden section search strategy to perform online identification and output resonance characteristic parameters. The notch filter unit is a dynamically adjustable notch filter connected to the resonance identification unit. Based on the resonance characteristic parameters, it generates filter parameters using the pre-distortion bilinear transform method, which are used to filter the torque command output by the speed loop. The parameter update unit is connected to the resonance identification unit and uses the sliding window Goertzel algorithm to monitor the changes in the resonance characteristic parameters in real time and dynamically update the core filter parameters of the notch filter unit online.

2. The large inertia spatial two-dimensional turntable resonance suppression control system according to claim 1, characterized in that, The sampling frequency of the current loop is higher than that of the speed loop; the sampling frequency of the speed loop is higher than that of the position loop; the speed acquisition unit, the resonance identification unit, the notch filter unit, and the parameter update unit operate synchronously with the speed loop and do not interfere with the positioning logic of the position loop and the response characteristics of the current loop.

3. The large inertia spatial two-dimensional turntable resonance suppression control system according to claim 1, characterized in that, The control system is applied to the ATP servo control system of a spaceborne large-inertia two-dimensional photoelectric turntable; the servo three-ring control unit, the speed acquisition unit, the resonance identification unit, the notch filter unit, and the parameter update unit are integrated into a programmable digital logic device and implemented in parallel using hardware logic circuits; the system is suitable for dual-axis servo control and resonance suppression of the positioning and pitch axes.

4. A method for suppressing resonance in a large-inertia two-dimensional spatial turntable based on a Goertzel notch filter, applied to the resonance suppression control system of a large-inertia two-dimensional spatial turntable as described in any one of claims 1 to 3, characterized in that, Includes the following steps: Step S100: Acquire the time-domain signal of the rotational speed fed back in real time by the rotational speed acquisition unit; Step S200: By improving the Goertzel algorithm and combining it with an adaptive threshold golden section search strategy, the rotational speed time-domain signal is identified online, and the resonance characteristic parameters are output. Step S300: Based on the resonant characteristic parameters, the core filter parameters are dynamically generated using the pre-distortion bilinear transform method to construct the notch filter; Step S400: The notch filter is embedded in series into the control path of the speed loop to filter the torque command output by the speed loop; In step S500, the Goertzel sliding window algorithm is used to monitor the changing trend of the resonant characteristic parameters in real time; when the changing trend exceeds a preset threshold, the process returns to step S200 to re-identify online in order to update the core filter parameters.

5. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In step S200, the improved Goertzel algorithm employs a second-order IIR recursive structure; the second-order IIR recursive structure calculates the amplitude and phase only for the target resonant frequency point in the rotational speed time-domain signal.

6. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In step S200, the adaptive threshold golden section search strategy includes: Step S201: Use second-order polynomial fitting to remove the trend term in the rotational speed time domain signal, and estimate the noise intensity by median absolute deviation to establish a noise benchmark; Step S202: Based on the noise reference and the global measured signal-to-noise ratio of the rotational speed time domain signal, dynamically and adaptively update the threshold coefficient to set a judgment threshold for distinguishing between reliable signals and noise. Step S203: Iteratively search for the resonance interval based on the judgment threshold and the golden ratio, and use two levels of sampling points, coarse sampling or fine sampling, depending on the current interval type; use coarse sampling to speed up in the noise interval, and use fine sampling to maintain accuracy in the resonance interval. Step S204: Using a weak decision counting and multi-window collaboration mechanism, multi-window joint determination is performed based on the results of the coarse sampling and the fine sampling to achieve global optimal identification of the resonant frequency and output the resonant characteristic parameters.

7. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In step S300, the pre-distortion bilinear transform method first corrects the resonant frequency shift through pre-distortion compensation, and then completes the filter discretization through bilinear transform to ensure that the core filtering parameters of the discretized notch filter are consistent with those in the continuous domain.

8. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In steps S200 and S300, the resonant characteristic parameters include the resonant center frequency, resonant amplitude, and resonant bandwidth; the core filtering parameters of the notch filter include the notch center frequency, notch depth, and notch width.

9. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In step S500, the sliding window Goertzel algorithm includes: Step S501: Set a sliding window of fixed length to buffer the rotational speed time-domain signal collected in multiple consecutive control cycles; Step S502: Update the rotational speed time domain signal in the sliding window in real time, and remove the historically outdated rotational speed time domain signal; Step S503: Based on the historical rotational speed time-domain signal sequence in the sliding window, the Goertzel algorithm is used to calculate the resonance characteristic parameters at each historical moment and generate the resonance parameter change trend. Step S504: Based on the trend of the resonant parameters, if the trend exceeds the preset threshold, proceed to step S200 to re-identify online; otherwise, proceed to step S501 to continue monitoring.

10. The method for suppressing resonance of a large-inertia two-dimensional turntable according to claim 4, characterized in that, In step S500, the preset threshold is dynamically set based on the global measurement signal-to-noise ratio of the rotational speed time-domain signal.

Images