Delay compensation filtering method for suppressing high-frequency measurement noise and related device

Abstract

Disclosed in the present invention are a delay compensation filtering method for suppressing high-frequency measurement noise and a related device. The method comprises: obtaining a real signal having time delay and amplitude distortion acquired by an inertial measurement unit on a quad-rotor unmanned aerial vehicle, and performing time delay compensation processing on the real signal to generate an initial filtering result; using an iterative filter having an angle truncation function to perform iterative optimization on the basis of the initial filtering result, and introducing an angle-limited constraint to truncate a signal having a large angle variation to maintain a smooth trajectory; and the output of the iterative filter being subject to boundary constraints centered on the initial filtering result, and if a filtering result output by the iterative filter exceeds a boundary, triggering a boundary processing mechanism to limit the filtering result within the boundary to obtain an underlying low-frequency real signal. The present invention effectively improves the signal quality, reduces the negative impact of delay caused by delay of a linear low-pass filter, and improves the smoothness and accuracy of signal processing and data transmission.

Description

A delay compensation filtering method and related equipment for suppressing high-frequency measurement noise Technical Field

[0001] The present invention relates to the field of signal processing technology, and in particular to a delay compensation filtering method, system, terminal and computer-readable storage medium for suppressing high-frequency measurement noise. Background Art

[0002] Delay-compensating filtering is a widely used method in signal processing. In practical applications, commonly used linear low-pass filtering of signals often results in delays, which can cause signal distortion or data inaccuracy. Delay-compensating filtering has been developed to address this issue. The basic principle of delay-compensating filtering is to approximate the signal's first-order Taylor expansion by introducing a differential, allowing the signal to compensate for phase and amplitude distortion after passing through the filter. This method can be implemented using various algorithms, including weighted averaging or interpolation techniques to estimate the delay value and adjust the signal accordingly. Delay-compensating filtering has been widely used in various fields, such as audio signal processing, video transmission, and communication systems.

[0003] An inertial measurement unit (IMU) is primarily composed of accelerometers and gyroscopes, and some also include magnetometers. Based on Newtonian classical mechanics, it is able to continuously output inertial information. With the development of micro-electromechanical system (MEMS) accelerometers and gyroscopes, IMUs have gradually become a research hotspot due to their advantages such as small size, light weight, low power consumption, low cost, and shock resistance. In the field of robotic navigation and control, IMUs are used to measure the angular velocity and acceleration of quadrotors. Many advanced controllers can estimate system disturbances by measuring acceleration, thereby improving control accuracy. However, IMUs are susceptible to factors such as measurement noise, low stability, temperature, and magnetic field interference, and their sensitivity to constraints varies significantly, especially when the quadrotor is rotating. These changes are more pronounced.

[0004] Linear filters are widely used to mitigate sensor measurement noise due to their simplicity, ease of use, and tunability. Low-pass filters effectively attenuate measurement noise and are independent of the statistical noise distribution. However, low-pass filters can introduce amplitude loss proportional to the signal frequency and significant phase lag before and after the cutoff frequency, leading to severe tracking errors. The development of fractional-order filters can achieve high bandwidth, but they still introduce latency into the system, similar to integer-order filters. In summary, linear filters can face the challenges of phase lag and reduced bandwidth due to amplitude loss in many applications. Although some zero-phase filters have been developed, they are primarily suitable for offline filtering, where future data is available. To achieve a complete system state with acceptable time delay, sliding mode techniques have been widely developed in recent years. The advantage of sliding mode filters is their ability to converge to a constant input within a finite time. However, the development of sliding mode techniques requires the knowledge of the Lipschitz constant (the upper bound of the nth-order derivative of the signal). Furthermore, if the measured signal contains random measurement noise, overshoot and chatter in the estimated state can be significantly amplified due to the sliding mode algorithm's sensitivity to noise and time step size. Adaptive time-delay compensation filters reconstruct the system state from the measured signal and use Taylor's first-order expansion to approximately compensate for phase delay and amplitude distortion. However, the differential operation introduced in the compensation process increases the filter's sensitivity to noise.

[0005] Sensor measurements are often contaminated by noise, and MEMS and IMUs are extremely sensitive to noise variations, especially in quadrotors, where the high-speed rotation of the propellers introduces significant noise into the IMU's measurements. Consequently, the IMU's raw measurement signal requires a low-pass filter with a large time constant to suppress high-frequency noise, but this introduces significant latency.

[0006] Therefore, the existing technology still needs to be improved and developed.

[0007] Summary of the Invention

[0008] The main purpose of the present invention is to provide a delay compensation filtering method, system, terminal and computer-readable storage medium for suppressing high-frequency measurement noise, aiming to solve the problems in the prior art where inertial measurement units are affected by measurement noise, low stability, temperature and magnetic field interference, signal delay, and significant changes in constrained sensitivity when measuring data.

[0009] To achieve the above object, the present invention provides a delay compensation filtering method for suppressing high-frequency measurement noise, the delay compensation filtering method for suppressing high-frequency measurement noise comprising the following steps:

[0010] Obtaining an original high-frequency signal collected by an inertial measurement unit on a quadrotor drone, performing a sliding average filter on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, performing time delay compensation processing on the low-frequency signal, and generating an initial filtering result;

[0011] An iterative filter with an angle truncation function is used to perform iterative optimization based on the initial filtering result, and an angle change constraint is introduced within adjacent sampling time intervals to maintain the smoothness of the trajectory;

[0012] A boundary constraint is added to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, a boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with low frequency and high signal-to-noise ratio.

[0013] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise, wherein the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone is obtained, the original high-frequency signal is subjected to sliding average filtering to obtain a low-frequency signal with time delay and amplitude distortion, the low-frequency signal is subjected to time delay compensation processing, and an initial filtering result is generated, specifically comprising:

[0014] The original high-frequency signal collected by the inertial measurement unit on the quadrotor drone is obtained, and a sliding average filter is performed on the original high-frequency signal to obtain a real signal with time delay and amplitude distortion. The low-frequency signal is expanded using a first-order Taylor method, and the compensated moving average filter in discrete form is designed as follows:

[0015] in, represents the output of the sliding average filter delay compensation, represents the differential signal of the moving average filter, represents the differential signal of the moving average filter, Δt represents the time delay, and k represents the time.

[0016] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise further comprises:

[0017] At time k-1, the estimated rate of change per unit time satisfies:

[0018] in, represents the estimated rate of change of the iterative filter at time k-1, represents the iterative filter estimate at time k-1, represents the iterative filter estimate at time k-2.

[0019] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise, wherein the iterative filter with an angle truncation function is used, iterative optimization is performed based on the initial filtering result, and an angle change constraint within adjacent sampling time intervals is introduced to maintain the smoothness of the trajectory, specifically comprising:

[0020] If the slope is the tangent of the angle β, it is expressed as tan(β);

[0021] At time k, the estimated result obtained by delay compensation filtering is interpolated with the estimated result at time k-1 to calculate the slope, which is used to calculate the rate of change of the iterative filter output:

[0022] in, represents the tangent of the angle α, expressed as tan(α), represents the estimated result obtained by the initial filter compensation calculation at time k;

[0023] The straight line at time k and the straight line at time k-1 intersect to form an angle γ. According to the relationship between trigonometric functions, the relative slope tan(γ[k]) is expressed as follows:

[0024] Among them, γ∈(-90°,90°), α[k] represents the rate of change of the iterative filter output, and β[k] represents the estimated rate of change of the iterative filter at time k-1. represents the estimated slope at time k-1;

[0025] The change in slope is affected by the known constant tan(γ max ) is limited so that tan(γ[k])<tan(γ max );

[0026] If the relative slope exceeds a given threshold tan(γ max ), then the noise dominates the result, and the current point is modified to ensure that the slope change is bounded;

[0027] Design a new relative slope And the following relationship is satisfied:

[0028] According to the geometric relationship of the updated slope, formula (5) can be rewritten as:

[0029] in, and Represents the update slope in two different directions.

[0030] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise further comprises:

[0031] According to the estimated slope, the upper and lower bounds of the update results are and

[0032] when When the current slope exceeds the lower bound, the slope is modified to

[0033] when When the current slope exceeds the upper limit, the slope is modified to

[0034] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise further comprises:

[0035] The final estimate is expressed in terms of the estimated slope as:

[0036] According to the designed angle constraint, The result is always smooth and can be described mathematically as:

[0037] At the same threshold tan(γ max ), if a threshold is given, then The size of the difference depends on and size.

[0038] Optionally, the delay compensation filtering method for suppressing high-frequency measurement noise, wherein a boundary constraint is added to the output of the iterative filter, and if the filtering result output by the iterative filter exceeds the boundary, a boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with a low frequency and a high signal-to-noise ratio, specifically comprising:

[0039] Boundary constraints are introduced into the output of time delay compensation, and filtering is performed. The filtered output is obtained by averaging two adjacent points. The filtered output is regarded as a middle line, which is expressed as:

[0040] The upper and lower boundaries are obtained based on the obtained middle line:

[0041] Among them, ε max is a given threshold;

[0042] The boundary defined based on time delay compensation provides a constraint for the iterative filtering process. If the iterative result exceeds the boundary, a boundary processing mechanism is designed to replace the current result into the boundary and continue the integration from the current point. The mathematical expression of the boundary processing mechanism is as follows:

[0043] Among them, the function min(a,b) is used to return the minimum value between two numbers, and the function max(a,b) is used to return the maximum value between two numbers;

[0044] The estimated value is processed by the boundary processing mechanism Always keep within the range [f min ,f max ] to ensure that the trajectory is within the predefined boundaries;

[0045] Generates smoothed estimates of the signal within the region where it is not bounded by boundaries The degree of smoothness is determined by formula (8).

[0046] In addition, to achieve the above-mentioned object, the present invention further provides a delay compensation filtering system for suppressing high-frequency measurement noise, wherein the delay compensation filtering system for suppressing high-frequency measurement noise comprises:

[0047] A time delay compensation module is used to obtain the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone, perform sliding average filtering on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, perform time delay compensation processing on the low-frequency signal, and generate an initial filtering result;

[0048] An angle truncation optimization module, configured to use an iterative filter with an angle truncation function to perform iterative optimization based on the initial filtering result and introduce angle change constraints within adjacent sampling time intervals to maintain the smoothness of the trajectory;

[0049] The boundary constraint module is used to add a boundary constraint to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, the boundary processing mechanism is triggered to limit the filtering result within the boundary to obtain a smooth signal with low frequency and high signal-to-noise ratio.

[0050] In addition, to achieve the above-mentioned objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a delay compensation filter program for suppressing high-frequency measurement noise stored in the memory and executable on the processor, wherein the delay compensation filter program for suppressing high-frequency measurement noise, when executed by the processor, implements the steps of the delay compensation filter method for suppressing high-frequency measurement noise as described above.

[0051] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a delay compensation filter program for suppressing high-frequency measurement noise, and when the delay compensation filter program for suppressing high-frequency measurement noise is executed by a processor, the steps of the delay compensation filter method for suppressing high-frequency measurement noise as described above are implemented.

[0052] In the present invention, an original high-frequency signal collected by an inertial measurement unit on a quadrotor drone is obtained, a sliding average filter is performed on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, and the low-frequency signal is subjected to time delay compensation processing to generate an initial filtering result. An iterative filter with an angle truncation function is used to perform iterative optimization based on the initial filtering result, and an angle change constraint within adjacent sampling time intervals is introduced to maintain the smoothness of the trajectory. A boundary constraint is added to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, a boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with a low frequency and high signal-to-noise ratio. The present invention can effectively improve signal quality, reduce the negative impact of delay caused by the delay of a linear low-pass filter, and provide more accurate and reliable data analysis. By using delay compensation filtering, angle truncation, and boundary constraints, the smoothness and accuracy of signal processing and data transmission can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] FIG1 is a flow chart of a preferred embodiment of a delay compensation filtering method for suppressing high-frequency measurement noise according to the present invention;

[0054] 2 is a schematic diagram of the definition of slope and angle in a preferred embodiment of the delay compensation filtering method for suppressing high-frequency measurement noise of the present invention;

[0055] 3 is a schematic diagram showing update slopes in two different directions in a preferred embodiment of the delay compensation filtering method for suppressing high-frequency measurement noise of the present invention;

[0056] 4 is a flow chart showing the basic principles of an iterative filtering algorithm with angle truncation in a preferred embodiment of the delay compensation filtering method for suppressing high-frequency measurement noise of the present invention;

[0057] 5 is a schematic diagram of the middle line and the boundary in a preferred embodiment of the delay compensation filtering method for suppressing high-frequency measurement noise of the present invention;

[0058] 6 is a discrete flow chart of a delay compensation filtering system with angle and boundary dual constraints in a preferred embodiment of the delay compensation filtering method for suppressing high-frequency measurement noise of the present invention;

[0059] 7 is a schematic diagram showing the principle of a preferred embodiment of a delay compensation filtering system for suppressing high-frequency measurement noise according to the present invention;

[0060] FIG8 is a schematic diagram of an operating environment of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0062] The delay compensation filtering method for suppressing high-frequency measurement noise according to a preferred embodiment of the present invention is shown in FIG1 . The delay compensation filtering method for suppressing high-frequency measurement noise comprises the following steps:

[0063] Step S10: obtaining an original high-frequency signal collected by an inertial measurement unit on a quadrotor drone, performing a sliding average filter on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, performing time delay compensation processing on the low-frequency signal, and generating an initial filtering result;

[0064] Step S20: using an iterative filter with an angle truncation function, performing iterative optimization based on the initial filtering result, and introducing an angle change constraint within adjacent sampling time intervals to maintain the smoothness of the trajectory;

[0065] Step S30: adding a boundary constraint to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, a boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with low frequency and high signal-to-noise ratio.

[0066] The present invention uses a delay compensation strategy based on Taylor expansion to alleviate the problems of time lag and amplitude distortion caused by the linear sliding average filter. This compensation step effectively alleviates the inherent time delay in the system and provides an initial estimate of the filtered signal. Based on the output of the time delay compensation, an iterative filter with angle constraints is proposed to ensure the smoothness of the filtering result. The iterative filter iteratively improves the estimate based on the previous estimate and the current input, and combines the angle constraints to achieve smoothness of the output result. In order to prevent potential instability or even divergence in the iterative filter, boundary constraints are introduced. These boundary constraints are based on the output of the time delay compensation and serve as limits on the filtering results. Simulation results and actual signals verify the effectiveness of the iterative filter.

[0067] Low-pass filters can effectively process some high-frequency noise and are not easily affected by statistical noise distribution. A moving average filter is a special case of a low-pass filter. Low-pass filters are mainly used to retain the low-frequency components of the signal by filtering out high-frequency components, thereby smoothing the signal and eliminating noise. The moving average filter smoothes the signal by calculating the average value of the data over a period of time. In this process, it uses a window (which can be fixed or variable size), and the data within the window is used to calculate the average value. The window size of the moving average filter determines its cutoff frequency, that is, it will smooth out high-frequency signal components. The moving average filter can be regarded as a low-pass filter, which achieves signal smoothing and noise elimination by averaging data. It limits the bandwidth of the signal to a certain extent, thus having the effect of a low-pass filter.

[0068] Although moving average filters are easy to use and adjust, they can easily cause amplitude loss proportional to the signal frequency, and time delay may appear in the filtered result. To solve this problem, a strategy needs to be designed to compensate for time delay and amplitude distortion. Therefore, the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone is obtained, and the original high-frequency signal is subjected to a sliding average filter to obtain a low-frequency signal with time delay and amplitude distortion. Using a first-order Taylor expansion to approximate the real signal, the compensated moving average filter in discrete form is designed as:

[0069] in, represents the output of the sliding average filter delay compensation, represents the differential signal of the moving average filter, represents the differential signal of the moving average filter, Δt represents the time delay, and k represents the time.

[0070] The application of the above-mentioned filter compensation strategy can reduce the phase delay. However, this strategy will introduce additional noise because it involves differential operations. Therefore, it is necessary to redesign the filter on this basis to eliminate and alleviate the noise impact. In reality, within a given sensor sampling period, the aircraft steering angle is bounded and smooth. In addition, the rate of change of velocity and acceleration is also bounded. If the measurement of any of these state variables changes suddenly and significantly, it will exceed the limitations of the system. This is due to external uncertainty, and the rate of change of these states must be limited. In other words, the estimated state at the current moment should be limited in its increase compared to the previous moment. To this end, the present invention designs an iterative filter (i.e., iterative filter) and uses angle as a constrained physical quantity to limit the signal output.

[0071] The input of the iterative angle limit filter is the output of the delay compensation. , and the filter output estimated at initialization strictly conforms to the basic calculus relationship. Figure 2 is a schematic diagram of the angle definition, with the horizontal axis representing the sampling points and the vertical axis representing the output results. The gray line represents the estimated slope at time k-1, and the black line represents the slope derived at time k.

[0072] At time k-1, the estimated rate of change per unit time satisfies:

[0073] in, represents the estimated rate of change of the iterative filter at time k-1, represents the iterative filter estimate at time k-1, represents the iterative filter estimate at time k-2. Geometrically, its rate of change is equal to its slope and is specific to an angle.

[0074] If the slope is the tangent of the angle β (ie, the mathematical expression of the slope is also called the tangent of the angle β), it is expressed as tan(β).

[0075] At time k, the estimated result obtained by delay compensation calculation is interpolated with the estimated result at time k-1 to calculate the slope, which is used to calculate the rate of change of the iterative filter output:

[0076] in, represents the tangent of the angle α, expressed as tan(α), Represents the estimated result obtained by the initial filter compensation calculation at time k.

[0077] The straight line at time k and the straight line at time k-1 intersect to form an angle γ. According to the relationship between trigonometric functions, the relative slope tan(γ[k]) can be deduced as follows:

[0078] Among them, γ∈(-90°,90°), α[k] represents the rate of change of the iterative filter output, and β[k] represents the estimated rate of change of the iterative filter at time k-1. represents the estimated slope at time k-1.

[0079] The change in slope is affected by the known constant tan(γ max ) is limited so that tan(γ[k])<tan(γ max ), and the value of this constant depends on the mobility of the system. The system performance can generally be obtained through preliminary experiments.

[0080] If the relative slope exceeds a given threshold tan(γ max ), then noise dominates the result, and the current point should be modified to ensure that the slope change is bounded.

[0081] In mathematical terms, we can design a new relative slope. And the following relationship is satisfied:

[0082] Therefore, the update slopes in two different directions can be expressed as shown in Figure 3 (the horizontal axis is the sampling point, and the vertical axis is the output result). As shown in Figure 3, tan(γ max ) is a threshold value, and the positive and negative values ​​correspond to the two gray lines in Figure 3, that is, clockwise or counterclockwise rotation along the gray circle.

[0083] According to the geometric relationship of the updated slope, formula (5) can be rewritten as:

[0084] in, and Represents the update slope in two different directions.

[0085] According to the estimated slope, the upper and lower bounds of the update results are and For example, when When the current slope exceeds the lower bound, the slope is modified to For example, when When the current slope exceeds the upper limit, the slope is modified to

[0086] The results of other different cases are shown in Figure 4. The final estimate is expressed as follows based on the estimated slope:

[0087] In general, according to the design angle constraints, The result is always smooth and can be described mathematically as:

[0088] From this, we can get some characteristics of the iterative filter: the iterative process takes into account the slope change between two adjacent time points; at the same threshold tan(γ max ), if a threshold is given, then The size of the difference depends on and size.

[0089] If the estimate is very large, then The change in will be very small, thus satisfying the given threshold tan(γ max In summary, the basic principle of the iterative filtering algorithm 1 with angle truncation from time k-1 to time k is shown in FIG4 .

[0090] In the iterative filter designed in the flowchart of Figure 4, an important parameter γ needs to be determined. max Only when this parameter takes a small value can the slope change rate of adjacent time intervals be guaranteed to be slow, thus obtaining a smoother trajectory. However, this will also lead to a longer response time for the system to track the filtering results, introducing a delay. In order to find a trade-off between smoothness and delay, a boundary constraint is introduced in the output of the time delay compensation, and filtering is performed. The filtered output is obtained by averaging two adjacent points. The filtered output is regarded as a middle line. From a mathematical perspective, the middle line is expressed as:

[0091] Based on the obtained middle line, the upper and lower boundaries can be deduced as:

[0092] Among them, ε max For a given threshold, a schematic diagram of the middle line and the boundaries is shown in FIG5 , which are schematic diagrams of the true signal, the upper boundary, the lower boundary, the middle line, and the delay compensation line, respectively.

[0093] The boundary defined based on time delay compensation provides a constraint for the iterative filtering process. If the iterative result exceeds the boundary, a boundary processing mechanism is designed to replace the current result into the boundary and continue the integration from the current point. The mathematical expression of the boundary processing mechanism is as follows:

[0094] Among them, the function min(a,b) is used to return the minimum value between two numbers, and the function max(a,b) is used to return the maximum value between two numbers; the estimated value is processed by the boundary processing mechanism. Always keep within the range [f min ,f max ] to ensure that the trajectory is within the predefined boundary. In addition, compared with the time delay compensation filter designed previously, the filter proposed here generates smooth estimates in the signal region that is not restricted by the boundary. This smooth trajectory estimation property is crucial because it prevents the iterative filter from diverging during the iteration process, thereby improving the overall performance of the filter.

[0095] As shown in Figure 6, the discrete flow chart of the delay-compensated filtering system with dual angle and boundary constraints proposed in this invention is shown. The proposed filtering process consists of three main components: the first component is time delay compensation, which is used to generate the initial filtering result. This step compensates for the system's inherent time delay and provides an initial estimate of the filtered signal. The second component is an iterative filter with an angle truncation function, based on the output of the time delay compensation, to ensure the smoothness of the filtering result. The iterative filter iteratively optimizes the estimate based on the previous estimate and the current input, while also introducing angle variation constraints within adjacent sampling time intervals to maintain the smoothness of the trajectory. The third component is to prevent potential instability in the iterative filter by introducing boundary constraints based on the output of the time delay compensation. These boundary constraints act as limits on the filtering result based on the output of the time delay compensation. If the output of the iterative filter exceeds the bounds, a mechanism is triggered to clamp the result within the bounds and continue the filtering process. Through these steps, if the bounds are not triggered, the final filtering result is smooth, effectively compensates for time delay, and maintains the desired signal estimate. Overall, the proposed filtering system depicted in Figure 6 combines time delay compensation, iterative filtering with angle shifting, and boundary constraints to achieve a smooth and stable estimate of the filtered signal.

[0096] This paper designs an iterative nonlinear filter with angle truncation, which significantly improves the smoothness of the filtered output and makes the filtered signal more stable by reducing high-frequency oscillations. Boundary constraints are added to the output of the iterative filter to prevent output distortion and divergence. The boundary mechanism helps to maintain the integrity and consistency of the filtered output. The filter parameters and their impact on performance are deeply studied, and an adjustment guide for the optimization results is provided. Through extensive simulation studies and analysis of actual signals, the effectiveness of the proposed method in different scenarios is verified.

[0097] This paper proposes a novel delay compensation and smoothing filter specifically for processing IMU measurement data in quadrotor aircraft. The proposed filter addresses challenges associated with propeller variations and high-frequency measurement noise, providing improved smoothness, reduced latency, and accurate state estimation. Comparative evaluations with existing filters, such as moving averages and previously studied delay compensation methods, highlight the proposed filter's ability to suppress high-frequency noise. Notably, the proposed filter effectively reduces signal distortion and latency, resulting in smoother and more accurate measurements. The effects of high-frequency noise can be effectively mitigated to ensure stable system performance. These advances are valuable for improving the measurement accuracy and responsiveness of IMUs in complex environments. Therefore, control and navigation systems can benefit from improved high-frequency noise filtering.

[0098] The present invention can be applied to the field of drones (the present invention is mainly applied to quad-rotor drones. It can effectively improve signal quality, reduce the negative impact of delay caused by linear low-pass filter delay, and provide more accurate and reliable data analysis. By using delay compensation filtering technology, the smoothness and accuracy of signal processing and data transmission can be improved. It is used to solve the delay problem in quad-rotor drone IMU measurement to improve the smoothness of the filter. The system integrates a delay compensation filter, an iterative filter with angle truncation and boundary constraints, thereby improving the accuracy of state estimation). It can also be applied to the field of robotics, and is also suitable for control, guidance and navigation and other fields.

[0099] Furthermore, as shown in FIG7 , based on the above-mentioned delay compensation filtering method for suppressing high-frequency measurement noise, the present invention also provides a delay compensation filtering system for suppressing high-frequency measurement noise, wherein the delay compensation filtering system for suppressing high-frequency measurement noise includes:

[0100] The time delay compensation module 51 is used to obtain the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone, perform a sliding average filter on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, perform time delay compensation processing on the low-frequency signal, and generate an initial filtering result;

[0101] An angle truncation optimization module 52 is used to use an iterative filter with an angle truncation function to perform iterative optimization based on the initial filtering result, and introduce angle change constraints within adjacent sampling time intervals to maintain the smoothness of the trajectory;

[0102] The boundary constraint module 53 is used to add a boundary constraint to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, the boundary processing mechanism is triggered to limit the filtering result within the boundary to obtain a smooth signal with low frequency and high signal-to-noise ratio.

[0103] Furthermore, as shown in FIG8 , based on the above-described delay compensation filtering method and system for suppressing high-frequency measurement noise, the present invention also provides a terminal, comprising a processor 10, a memory 20, and a display 30. FIG8 illustrates only some components of the terminal, but it should be understood that implementation of all illustrated components is not required, and more or fewer components may be implemented instead.

[0104] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Furthermore, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a delay compensation filter program 40 for suppressing high-frequency measurement noise is stored on the memory 20, and the delay compensation filter program 40 for suppressing high-frequency measurement noise can be executed by the processor 10, thereby realizing the delay compensation filter method for suppressing high-frequency measurement noise in the present application.

[0105] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program code or process data stored in the memory 20, such as executing the delay compensation filtering method for suppressing high-frequency measurement noise.

[0106] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0107] In one embodiment, the steps of the delay compensation filtering method for suppressing high-frequency measurement noise are implemented when the processor 10 executes the delay compensation filtering program 40 for suppressing high-frequency measurement noise in the memory 20 .

[0108] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a delay compensation filter program for suppressing high-frequency measurement noise, and when the delay compensation filter program for suppressing high-frequency measurement noise is executed by a processor, the steps of the delay compensation filter method for suppressing high-frequency measurement noise as described above are implemented.

[0109] In summary, the present invention provides a delay compensation filtering method and related equipment for suppressing high-frequency measurement noise. The method includes: obtaining an original high-frequency signal collected by an inertial measurement unit on a quadrotor drone, performing a sliding average filter on the original high-frequency signal to obtain a low-frequency signal with time delay and amplitude distortion, performing time delay compensation processing on the low-frequency signal, and generating an initial filtering result; using an iterative filter with an angle truncation function, iteratively optimizing according to the initial filtering result, and introducing an angle change constraint within adjacent sampling time intervals to maintain the smoothness of the trajectory; adding a boundary constraint to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, a boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with a low frequency and high signal-to-noise ratio, and preventing the iterative filter from diverging during the iterative process. The present invention can effectively improve signal quality, reduce the negative impact of delay caused by the delay of a linear low-pass filter, and provide more accurate and reliable data analysis. By using delay compensation filtering, angle truncation and boundary constraints, the smoothness and accuracy of signal processing and data transmission can be improved.

[0110] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal comprising the element.

[0111] Of course, those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When the program is executed, it can include the processes in the above-described method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0112] It should be understood that the application of the present invention is not limited to the above examples. For those skilled in the art, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A delay compensation filtering method for suppressing high frequency measurement noise, characterized in that: The delay compensation filtering method for suppressing high-frequency measurement noise includes: Acquire the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone, perform sliding average filtering on the original high-frequency signal, obtain a low-frequency signal with time delay and amplitude distortion, perform time delay compensation processing on the low-frequency signal, and generate an initial filtering result; An iterative filter with an angle truncation function is used to perform iterative optimization according to the initial filtering result, and an angle change constraint is introduced within adjacent sampling time intervals to maintain the smoothness of the trajectory; A boundary constraint is added to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, the boundary processing mechanism is triggered to limit the filtering result within the boundary, thereby obtaining a smooth signal with a low frequency and a high signal-to-noise ratio.

2. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 1, characterized in that: The method of obtaining an original high-frequency signal collected by an inertial measurement unit on a quad-rotor drone, performing a sliding average filter on the original high-frequency signal, obtaining a low-frequency signal with time delay and amplitude distortion, performing time delay compensation processing on the low-frequency signal, and generating an initial filtering result specifically includes: The original high-frequency signal collected by the inertial measurement unit on the quadrotor drone is obtained, and the original high-frequency signal is subjected to sliding average filtering to obtain a real signal with time delay and amplitude distortion. The low-frequency signal is expanded using a first-order Taylor, and the compensated moving average filter in discrete form is designed as follows: in, represents the output of the moving average filter delay compensation, represents the differential signal of the moving average filter, represents the differential signal of the moving average filter, Δt represents the time delay, and k represents the time.

3. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 2, characterized in that: The delay compensation filtering method for suppressing high-frequency measurement noise also includes: At time k-1, the estimated rate of change per unit time satisfies: in, represents the estimated rate of change of the iterative filter at time k-1, represents the iterative filter estimate at time k-1, represents the iterative filter estimate at time k-2.

4. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 3, characterized in that: The iterative filter with angle truncation function is used to perform iterative optimization according to the initial filtering result, and an angle change constraint is introduced within adjacent sampling time intervals to maintain the smoothness of the trajectory, specifically including: If the slope is the tangent of the angle β, it is expressed as tan(β); At time k, the estimated result obtained by delay compensation filtering is interpolated with the estimated result at time k-1 to calculate the slope, which is used to calculate the rate of change of the iterative filter output: in, represents the tangent of the angle α, expressed as tan(α), represents the estimated result obtained by the initial filter compensation calculation at time k; The straight line at time k and the straight line at time k-1 intersect to form an angle γ. According to the relationship between trigonometric functions, the relative slope tan(γ[k]) is expressed mathematically as follows: Among them, γ∈(-90°,90°), α[k] represents the rate of change of the iterative filter output, and β[k] represents the estimated rate of change of the iterative filter at time k-1. represents the estimated slope at time k-1; The change in slope is governed by the known constant tan(γ max ) is limited so that tan(γ[k])<tan(γ max ); If the relative slope exceeds a given threshold tan(γ max ), then the noise dominates the result and the current point is modified to satisfy the bounded slope change; Design a new relative slope And the following relationship is satisfied: According to the geometric relationship of the updated slope, formula (5) is rewritten as: in, and Represents the update slope in two different directions.

5. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 4, characterized in that: The delay compensation filtering method for suppressing high-frequency measurement noise also includes: According to the estimated slope, the upper and lower bounds of the update results are and when , the current slope exceeds the lower bound and the slope is modified to when , the current slope exceeds the upper limit, and the slope is modified to 6. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 5, characterized in that: The delay compensation filtering method for suppressing high-frequency measurement noise also includes: The final estimate is expressed in terms of the estimated slope as: According to the designed angle constraints, The result is always smooth and can be described mathematically as: At the same threshold tan(γ max ), if a threshold is given, then The size of the difference depends on and size.

7. The delay compensation filtering method for suppressing high-frequency measurement noise according to claim 6, characterized in that: The step of adding a boundary constraint to the output of the iterative filter, and triggering a boundary processing mechanism to limit the filtering result within the boundary if the filtering result output by the iterative filter exceeds the boundary, thereby obtaining a smooth signal with a low frequency and a high signal-to-noise ratio, specifically includes: Boundary constraints are introduced into the output of time delay compensation, and filtering is performed. The filtered output is obtained by averaging two adjacent points, where the filtered output is regarded as a middle line, which is expressed as: The upper and lower boundaries are obtained based on the obtained middle line: Among them, ε max is a given threshold; The boundary defined based on time delay compensation provides constraints for the iterative filtering process. If the iterative result exceeds the boundary, a boundary processing mechanism is designed to substitute the current result into the boundary and continue to integrate from the current point. The mathematical expression of the boundary processing mechanism is as follows: Among them, the function min(a,b) is used to return the minimum value between two numbers, and the function max(a,b) is used to return the maximum value between two numbers; The estimated value is processed by the boundary processing mechanism Always keep within the range [f min ,f max ] to ensure that the trajectory is within the predefined boundaries; Generates smoothed estimates of the region where the signal is not bounded by boundaries The degree of smoothness is determined by formula (8).

8. A delay compensation filtering system for suppressing high frequency measurement noise, characterized in that: The delay compensation filtering system for suppressing high-frequency measurement noise includes: The time delay compensation module is used to obtain the original high-frequency signal collected by the inertial measurement unit on the quadrotor drone, perform sliding average filtering on the original high-frequency signal, obtain a low-frequency signal with time delay and amplitude distortion, perform time delay compensation processing on the low-frequency signal, and generate an initial filtering result; An angle truncation optimization module, used to use an iterative filter with an angle truncation function, perform iterative optimization according to the initial filtering result, and introduce angle change constraints within adjacent sampling time intervals to maintain the smoothness of the trajectory; The boundary constraint module is used to add boundary constraints to the output of the iterative filter. If the filtering result output by the iterative filter exceeds the boundary, the boundary processing mechanism is triggered to limit the filtering result within the boundary to obtain a smooth signal with low frequency and high signal-to-noise ratio.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a delay compensation filter program for suppressing high-frequency measurement noise stored in the memory and executable on the processor. When the delay compensation filter program for suppressing high-frequency measurement noise is executed by the processor, the steps of the delay compensation filter method for suppressing high-frequency measurement noise as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a delay compensation filter program for suppressing high-frequency measurement noise. When the delay compensation filter program for suppressing high-frequency measurement noise is executed by a processor, the steps of the delay compensation filter method for suppressing high-frequency measurement noise as described in any one of claims 1 to 7 are implemented.

Images