A pan-less dynamic compensation water flow rate measurement system for vibration environments

Abstract

The present invention discloses a gimbal-free dynamic compensation water flow velocity measurement system for use in a vibrating environment. The system comprises a multi-source sensor module, a dynamic motion compensation module, an adaptive Kalman filter group, and a data credibility assessment module. The multi-source sensor module is used to respectively collect platform attitude information, original water flow velocity, and platform height. The dynamic motion compensation module comprises an attitude solver module and an error calculation submodule to achieve dynamic flow velocity compensation in a gimbal-free environment. The adaptive Kalman filter group comprises an angular velocity filter and a flow velocity filter. The angular velocity filter dynamically adjusts filter parameters based on the intensity of the platform's angular motion, and the flow velocity filter dynamically switches processing strategies based on the validity of the water flow signal measured by radar. The data credibility assessment module is used to perform stability assessment based on a sliding window variance-trend joint detection algorithm. The present invention integrates attitude solving and motion compensation to achieve accurate water flow velocity measurement on a gimbal-free platform in a vibrating environment.

Description

Technical Field

[0001] The present invention relates to the technical field of water flow rate measurement, and in particular to a pan-less dynamic compensation water flow rate measurement system for use in a vibration environment. Background Art

[0002] Real-time measurement of water velocity has become a key sensing tool in a variety of fields, including fluid environment monitoring, inland water surveys, urban flood control, and water resource management. Traditional methods for measuring water velocity rely on equipment installed on stationary platforms, such as radar velocimeters and ultrasonic flowmeters deployed on fixed bridges, embankments, or ground supports. While these measurement methods can provide relatively accurate and stable data in static environments, they are subject to numerous limitations when used on dynamic platforms.

[0003] With the widespread use of mobile platforms such as drones and unmanned boats, hydrological monitoring based on dynamic platforms has become a trend. However, when velocity sensors are mounted on such platforms, the platform's dynamic attitude inevitably changes in pitch, roll, and yaw during motion, introducing additional errors into the measured velocity data. For example, when using millimeter-wave radar to measure surface velocity, the platform's rotation can cause the beam direction to deviate from the target, or even measure velocity components generated by the platform's own motion. Furthermore, high-frequency interference generated by platform vibration can be superimposed on the velocity signal, compromising the stability and reliability of the original measurement data.

[0004] To reduce the interference of platform motion on measurement results, existing research mainly uses attitude compensation methods for correction, such as using the angle measured by the IMU for static compensation. However, such methods usually assume that the platform attitude changes slowly or regularly, and are difficult to adapt to the rapid and complex dynamic attitude changes in real environments. At the same time, existing filtering methods such as simple low-pass filtering or fixed-parameter Kalman filters often have problems such as response lag and signal distortion when dealing with intense motion or high-noise inputs. In addition, the credibility assessment of raw water flow data is relatively weak, and there is a lack of joint analysis of data stability and trend, resulting in subsequent hydrological modeling and control strategies relying on unreliable data, which poses potential risks.

[0005] In particular, in current unmanned platform applications, traditional mechanical gimbal systems are often eliminated to save weight and energy and simplify structural design, making it impossible for sensors such as radar to maintain a constant attitude. These gimbal-free applications exacerbate the dynamic coupling between sensor and platform attitude, further increasing the challenges posed by attitude resolution, motion compensation, and data filtering technologies.

[0006] Therefore, how to provide a pan-tilt-free dynamic compensation water flow rate measurement system for use in a vibration environment is an urgent problem that those skilled in the art need to solve. Summary of the Invention

[0007] One objective of the present invention is to provide a gimbal-free dynamic compensation water velocity measurement system for use in vibrating environments. This system integrates attitude calculation and error compensation, combined with adaptive filtering and a sliding window credibility assessment mechanism, to achieve high-precision dynamic measurement of water velocity in vibrating environments on a gimbal-free platform. The system offers advantages such as strong real-time performance, robust disturbance resistance, and compatibility with complex platforms. It addresses existing issues such as insufficient dynamic compensation accuracy, delayed filtering response, and difficulty assessing data reliability, making it suitable for mobile measurement scenarios such as drones and unmanned vessels.

[0008] According to an embodiment of the present invention, a pan-tilt-less dynamic compensation water flow rate measurement system for use in a vibration environment includes a multi-source sensor module, a dynamic motion compensation module, an adaptive Kalman filter group, and a data credibility assessment module;

[0009] The multi-source sensor module includes a three-axis MEMS inertial measurement unit, a millimeter-wave radar flow velocity sensor, and an ultrasonic altimeter, which are used to respectively collect platform attitude information, water flow original velocity, and platform height;

[0010] The dynamic motion compensation module includes an attitude solver module and an error calculation submodule. The attitude solver module is used to perform attitude calculation based on the platform attitude information and estimate the attitude change of the platform in a vibrating state in real time. The error calculation submodule removes the error velocity component introduced by the platform's own motion from the original water flow velocity based on the platform attitude information and angular velocity changes, thereby achieving dynamic flow velocity compensation without a pan-tilt structure.

[0011] The adaptive Kalman filter group includes an angular velocity filter and a flow rate filter. The angular velocity filter dynamically adjusts the filter parameters according to the intensity of the platform's angular motion. The flow rate filter dynamically switches the processing strategy according to the effectiveness of the water flow signal measured by the radar, performing different filtering processes on zero values, small signals, and valid signals respectively.

[0012] The data credibility assessment module is used to judge the stability of the water flow measurement data after dynamic compensation and filtering based on the sliding window variance-trend joint detection algorithm.

[0013] Optionally, the operating frequency of the millimeter wave radar flow velocity sensor is 24 GHz, the speed measurement range is 0.1 m / s to 20 m / s, and the measurement accuracy is 0.01 m / s.

[0014] Optionally, the attitude solver module uses a Mahony complementary filtering algorithm and a fourth-order Runge-Kutta integration method to solve the platform attitude information. The Mahony complementary filtering algorithm is based on the measurement information of the three-axis gyroscope and accelerometer, and realizes angular velocity correction through nonlinear error feedback, including:

[0015] set up For the platform at all times The attitude quaternion represents the rotation state of the platform from the inertial coordinate system to the body coordinate system. The initial value is the unit quaternion ;

[0016] Get the angular velocity vector measured by the three-axis gyroscope ,in They are the angular velocities around the X, Y, and Z axes, and the gravity direction vector measured by the accelerometer is obtained ;

[0017] Based on attitude quaternion Calculate the current estimated gravity direction vector :

[0018] ;

[0019] in, Quaternion representing attitude The conjugate quaternion of Represents quaternion multiplication operation;

[0020] The gravity direction vector measured by the accelerometer and the current estimated gravity direction vector Perform vector cross multiplication to obtain the gravity direction error vector ;

[0021] Constructing a corrected angular velocity vector with proportional-integral feedback :

[0022] ;

[0023] in, represents the proportional gain coefficient, represents the integral gain coefficient, Represents the gravity direction error vector The time integral of , and the corresponding correction angular velocity quaternion is constructed ;

[0024] The fourth-order Runge-Kutta integration method is used to numerically integrate the quaternion differential equation, and the integration time step is , the update process is:

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] The final updated attitude quaternion is:

[0030] ;

[0031] in, 、 、 and It represents the intermediate increment of Runge-Kutta integral, reflecting the trend of differential change. Represents the updated attitude quaternion.

[0032] Optionally, the error calculation submodule includes a radar installation parameter online calibration unit, a beam direction compensation unit, and a motion error speed compensation unit;

[0033] The radar installation parameter online calibration unit is used to correct the installation position parameters of the millimeter wave radar in the body coordinate system in real time based on the updated attitude quaternion output by the attitude solver module. The installation position parameter online calibration function satisfies:

[0034] ;

[0035] in, Represents the installation position vector of the millimeter-wave radar in the body coordinate system, and Indicates the nominal installation position of the millimeter-wave radar in the horizontal plane of the aircraft. and Indicates the online correction value of the horizontal installation deviation, Represents column vector form;

[0036] The beam direction compensation unit is based on the angle between the radar beam direction and the horizontal plane. , calculate the beam direction compensation factor :

[0037] ;

[0038] in, Represents the projection of the radar beam direction on the horizontal plane of the platform, which is obtained by rotating the static beam direction vector based on the attitude quaternion. represents the norm;

[0039] The motion error speed compensation unit calculates the error speed component of the platform's own rotational motion in three-dimensional space to the original water flow speed:

[0040] ;

[0041] in, represents the error velocity component, represents the rotation matrix from the body coordinate system to the inertial coordinate system, represents the angular velocity vector;

[0042] The error velocity component is removed from the original water flow velocity, thereby realizing dynamic compensation for the error introduced into the platform motion without a pan-tilt structure.

[0043] Optionally, the angular velocity filter adopts a two-stage series structure, including:

[0044] The first stage is a sliding average filter, which is used to perform preliminary smoothing on the raw angular velocity data to suppress high-frequency noise fluctuations. The width of the sliding average filter window is 5 to 15 data points.

[0045] The second level is A filter is used to further predict the angular velocity trend based on the sliding average filter result and enhance the response capability to dynamic changes, The filter parameters are dynamically adjusted according to the current motion state of the platform. The adjustment rules are as follows:

[0046] When the platform is in static mode, set ;

[0047] When the platform is in dynamic mode, set ;

[0048] in, represents the position update coefficient, represents the speed update coefficient, Represents the acceleration update coefficient;

[0049] The platform state is determined by the angular velocity amplitude. If the angular velocity is lower than the set amplitude threshold, it enters the static mode; otherwise, it enters the dynamic mode.

[0050] The two-stage filtering structure is combined to enhance the stability of the angular velocity signal and adjust the response performance under different vibration intensities, providing more reliable angular velocity input for the attitude solver module and the error calculation submodule.

[0051] Optionally, the flow rate filter includes a zero value processing unit, a small signal holding unit and a valid signal tracking unit;

[0052] The zero value processing unit is used to identify and eliminate invalid water flow signals that are continuously zero. When the continuous zero value exceeds the set continuous threshold, the processing is triggered. The continuous threshold setting range is 30 to 100 sampling points, which is used to eliminate the influence of false data caused by long-term no signal or echo failure;

[0053] The small signal holding unit is used to hold the small signal, with an activation threshold setting range of 0.05m / s to 0.2m / s. When the signal amplitude is detected to be lower than the activation threshold but persists, the small signal validity is automatically maintained;

[0054] The effective signal tracking unit is used to perform dynamic smoothing on the confirmed effective signal. The response speed is controlled by a time constant. The setting range of the time constant is 0.1s~1.0s, which is used to balance the response delay and the data fluctuation suppression capability.

[0055] Optionally, the sliding window variance-trend joint detection algorithm specifically includes:

[0056] The continuous water flow measurement data is divided into a sliding window sequence. The minimum number of stable points in each sliding window is set in the range of 10 to 50, and the sliding window step size is set in the range of 1 to 5 data points.

[0057] For the data within each sliding window, the coefficient of variation is calculated as a stability indicator:

[0058] ;

[0059] in, represents the coefficient of variation, represents the standard deviation of the data samples in the sliding window, Represents the average value of the data samples in the sliding window. When , the current sliding window is judged to be a stable interval;

[0060] The trend indicator is calculated using the first-order difference method to determine whether the data has a continuous change trend:

[0061] ;

[0062] in, Represents a trend indicator, represents the total number of sliding window data samples, Indicates the first data samples, Indicates the first data samples, represents a symbolic function, when , the current sliding window is determined to be an interval with no significant trend change.

[0063] The beneficial effects of the present invention are:

[0064] First, the present invention effectively improves the real-time and stability of attitude estimation in dynamic environments by introducing an attitude solution strategy that combines the Mahony complementary filter algorithm with the fourth-order Runge-Kutta integration method. Compared to traditional methods that rely solely on gyroscope integration or simple Kalman filtering, this solution structure utilizes accelerometers to correct long-term integration errors and reduces attitude fluctuations through high-order numerical integration, thereby accurately acquiring platform attitude information even under severe vibration or rapid flight conditions.

[0065] Secondly, regarding error compensation, this invention constructs an error velocity model based on attitude quaternions, angular velocity vectors, and radar installation parameters, dynamically eliminating non-fluid velocity components introduced by the platform's rotational motion. This allows for precise compensation of raw water velocity data without a gimbal. Furthermore, the system integrates online calibration of the radar installation position and beam direction projection compensation mechanisms, ensuring that the radar's measurement direction remains geometrically consistent with the actual water surface flow direction, fundamentally reducing measurement deviations caused by structural installation errors.

[0066] In addition, in terms of data processing, the present invention designs a set of adaptive Kalman filter groups, which configure two-stage filter structures according to the different characteristics of angular velocity and flow rate signals. The filter series structure automatically adjusts the filter parameters by detecting the platform's motion state, improving noise immunity and responsiveness under different dynamic conditions. The flow velocity filter has three processing units: zero-value processing, small signal retention, and effective signal tracking. These units can handle a variety of complex signal states, including still water, disturbed microflow, and normal flow, significantly improving the continuity and reliability of water velocity data.

[0067] Finally, to further ensure the credibility of the system's output data, the present invention constructed a sliding window variance-trend joint detection algorithm. By analyzing the coefficient of variation and first-order trend indicators of the measured data, it can accurately identify stable sections and non-trending areas of the data, giving the system the ability to quantitatively judge the credibility of the measurement results. This mechanism not only improves the system's ability to automatically identify low-quality data but also provides an important reliability basis for subsequent data fusion and hydrological modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0069] Figure 1This is a schematic structural diagram of the pan-less dynamic compensation water flow rate measurement system for use in a vibration environment proposed by the present invention;

[0070] Figure 2 This is a flowchart of the Mahony complementary filter and fourth-order Runge-Kutta integration of the attitude solver module of the pan-tilt-less dynamic compensation water flow rate measurement system in a vibration environment proposed by the present invention;

[0071] Figure 3 This is a motion compensation principle diagram in the error calculation submodule of the pan-tilt-free dynamic compensation water flow rate measurement system for a vibration environment proposed by the present invention. DETAILED DESCRIPTION

[0072] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0073] refer to Figure 1-Figure 3 , a pan-tilt-free dynamic compensation water flow rate measurement system for vibration environment, including a multi-source sensor module, a dynamic motion compensation module, an adaptive Kalman filter group and a data credibility assessment module;

[0074] The multi-source sensor module includes a three-axis MEMS inertial measurement unit, a millimeter-wave radar flow velocity sensor, and an ultrasonic altimeter, which are used to respectively collect platform attitude information, water flow original velocity, and platform height;

[0075] The dynamic motion compensation module includes an attitude solver module and an error calculation submodule. The attitude solver module is used to perform attitude calculation based on the platform attitude information and estimate the attitude change of the platform in a vibrating state in real time. The error calculation submodule removes the error velocity component introduced by the platform's own motion from the original water flow velocity based on the platform attitude information and angular velocity changes, thereby achieving dynamic flow velocity compensation without a pan-tilt structure.

[0076] The adaptive Kalman filter group includes an angular velocity filter and a flow rate filter. The angular velocity filter dynamically adjusts the filter parameters according to the intensity of the platform's angular motion. The flow rate filter dynamically switches the processing strategy according to the effectiveness of the water flow signal measured by the radar, performing different filtering processes on zero values, small signals, and valid signals respectively.

[0077] The data credibility assessment module is used to judge the stability of the water flow measurement data after dynamic compensation and filtering based on the sliding window variance-trend joint detection algorithm.

[0078] By constructing an integrated multi-module water flow velocity measurement architecture, this system achieves high-precision flow velocity sensing in vibrating environments and without a gimbal. The overall system integrates a multi-source sensor module, a dynamic motion compensation module, an adaptive Kalman filter bank, and a data credibility assessment module to form a closed data flow loop. Compared to traditional static platforms or structures that rely on mechanical gimbals, this system does not rely on external stabilization equipment. Instead, it uses algorithms to perform real-time sensing and compensation for platform posture and motion errors, adapting to the complex dynamic characteristics of the platform in non-stationary motion in the air or on the water. It is particularly suitable for lightweight platforms such as drones and unmanned boats, achieving a flow velocity monitoring solution that combines miniaturization, intelligence, and high robustness.

[0079] In this embodiment, the operating frequency of the millimeter wave radar flow velocity sensor is 24 GHz, the velocity measurement range is 0.1 m / s to 20 m / s, and the measurement accuracy is 0.01 m / s.

[0080] The millimeter-wave radar sensor operates at a frequency of 24 GHz and boasts strong anti-interference and penetration capabilities, maintaining stable velocity measurement even in complex water environments. Its velocity measurement range covers the common flow velocity range of natural water bodies, with an accuracy of 0.01 m / s, ensuring data sampling granularity and responsiveness to velocity changes. Using a radar unit of this specification enables the system to capture subtle flows and highly dynamic changes while maintaining measurement distance, meeting the needs of a variety of application scenarios, from steady water flows to turbulent currents. This improves the system's versatility and measurement accuracy, making it particularly suitable for deployment in complex external environments to perform long-term, continuous velocity observations.

[0081] The ultrasonic altimeter enables the platform to accurately obtain its relative height relative to the water surface while in flight or afloat. Its wide measurement range and high accuracy effectively adapt to monitoring platform altitude changes at varying flight altitudes or under choppy conditions, ensuring that radar velocity data accurately corresponds to the actual distance to the water surface. Altitude data, a crucial auxiliary variable for error compensation and attitude dynamics, effectively improves measurement accuracy and the integrity of the compensation model. The introduction of a high-frequency, high-precision altimeter effectively enhances the system's ability to detect platform attitude changes, improving the spatial resolution and geometric registration capabilities of velocity measurements.

[0082] In this embodiment, the attitude solver module uses the Mahony complementary filter algorithm and the fourth-order Runge-Kutta integration method to solve the platform attitude information. The Mahony complementary filter algorithm is based on the measurement information of the three-axis gyroscope and accelerometer, and realizes angular velocity correction through nonlinear error feedback, including:

[0083] set up For the platform at all times The attitude quaternion represents the rotation state of the platform from the inertial coordinate system to the body coordinate system. The initial value is the unit quaternion ;

[0084] Get the angular velocity vector measured by the three-axis gyroscope ,in They are the angular velocities around the X, Y, and Z axes, and the gravity direction vector measured by the accelerometer is obtained ,

[0085] Based on attitude quaternion Calculate the current estimated gravity direction vector :

[0086] ;

[0087] in, Quaternion representing attitude The conjugate quaternion of Represents quaternion multiplication operation;

[0088] The gravity direction vector measured by the accelerometer and the current estimated gravity direction vector Perform vector cross multiplication to obtain the gravity direction error vector ;

[0089] Constructing a corrected angular velocity vector with proportional-integral feedback :

[0090] ;

[0091] in, represents the proportional gain coefficient, represents the integral gain coefficient, Represents the gravity direction error vector The time integral of , and the corresponding correction angular velocity quaternion is constructed ;

[0092] The fourth-order Runge-Kutta integration method is used to numerically integrate the quaternion differential equation, and the integration time step is , the update process is:

[0093] ;

[0094] ;

[0095] ;

[0096] ;

[0097] The final updated attitude quaternion is:

[0098] ;

[0099] in, 、 、 and It represents the intermediate increment of Runge-Kutta integral, reflecting the trend of differential change. Represents the updated attitude quaternion.

[0100] By integrating the Mahony complementary filter algorithm and the fourth-order Runge-Kutta numerical integration method, the system can stably output high-precision attitude estimation results in highly vibrating or highly dynamic environments. The Mahony filter combines accelerometer and gyroscope data to correct angular velocity drift through nonlinear feedback, effectively combating long-term error accumulation in the inertial measurement unit. The Runge-Kutta integration method further improves the numerical stability and accuracy of the solution. This combined solution requires no external magnetometer support, boasts high computational efficiency and strong real-time performance, and is particularly suitable for resource-constrained platforms. Attitude information, the core basis for error compensation and radar beam angle correction, directly determines measurement accuracy. The high reliability of this module significantly enhances the overall system performance.

[0101] In this embodiment, the error calculation submodule includes a radar installation parameter online calibration unit, a beam direction compensation unit, and a motion error speed compensation unit;

[0102] The radar installation parameter online calibration unit is used to correct the installation position parameters of the millimeter wave radar in the body coordinate system in real time based on the updated attitude quaternion output by the attitude solver module. The installation position parameter online calibration function satisfies:

[0103] ;

[0104] in, Represents the installation position vector of the millimeter-wave radar in the body coordinate system, and Indicates the nominal installation position of the millimeter-wave radar in the horizontal plane of the aircraft. and Indicates the online correction value of the horizontal installation deviation, Represents column vector form;

[0105] The beam direction compensation unit is based on the angle between the radar beam direction and the horizontal plane. , calculate the beam direction compensation factor :

[0106] ;

[0107] in, Represents the projection of the radar beam direction on the horizontal plane of the platform, which is obtained by rotating the static beam direction vector based on the attitude quaternion. represents the norm;

[0108] The motion error speed compensation unit calculates the error speed component of the platform's own rotational motion in three-dimensional space to the original water flow speed:

[0109] ;

[0110] in, represents the error velocity component, represents the rotation matrix from the body coordinate system to the inertial coordinate system, represents the angular velocity vector;

[0111] The error velocity component is removed from the original water flow velocity, thereby realizing dynamic compensation for the error introduced into the platform motion without a pan-tilt structure.

[0112] The error compensation submodule accurately eliminates radar velocity errors caused by platform motion through multi-level modeling. First, the system supports an online calibration mechanism for the radar's mounting position, dynamically correcting actual position deviations and eliminating the impact of structural assembly errors on velocity measurement. Second, the beam direction compensation factor calculation method, combined with attitude solution results, effectively corrects for deviations in the radar's oblique beam in non-horizontal conditions. Third, a rotational motion model between angular velocity and the mounting vector is introduced to compensate for the additional velocity component generated by the platform's own rotation during radar perception.

[0113] In addition, it is worth noting that the platform rotation error velocity component and the projection deviation caused by the radar beam direction belong to two different sources of measurement errors, which are handled by different compensation mechanisms. The platform rotation error velocity component is calculated by the cross product relationship between the angular velocity and the radar installation position vector, combined with the rotation matrix generated by the attitude solution, and is used to compensate for the additional velocity of the radar measurement value caused by the rotation of the platform around the three-dimensional axis. The calculation result is a vector quantity that reflects the disturbance of the instantaneous motion state to the velocity measurement result. The beam direction compensation factor It is a geometric projection factor calculated based on the angle between the radar beam and the horizontal plane in the platform coordinate system. It is mainly used to correct the measurement deviation caused by the inconsistency between the radar speed measurement direction and the actual horizontal direction. This is a scalar compensation factor and doesn't directly contribute to the calculation of 3D velocity components. However, it can be used to further restore the true horizontal component of the radar-measured velocity during the velocity output phase after compensation, thereby improving the consistency between the measured data and the actual water flow state. The two work together to form a full-link error control mechanism, from dynamic disturbance compensation to directional correction, ensuring the system can still output high-precision water flow velocity data stably without a gimbal.

[0114] In this embodiment, the angular velocity filter adopts a two-stage series structure, including:

[0115] The first stage is a sliding average filter, which is used to perform preliminary smoothing on the raw angular velocity data to suppress high-frequency noise fluctuations. The width of the sliding average filter window is 5 to 15 data points.

[0116] The second level is A filter is used to further predict the angular velocity trend based on the sliding average filter result and enhance the response capability to dynamic changes, The filter parameters are dynamically adjusted according to the current motion state of the platform. The adjustment rules are as follows:

[0117] When the platform is in static mode, set ;

[0118] When the platform is in dynamic mode, set ;

[0119] in, represents the position update coefficient, represents the speed update coefficient, Represents the acceleration update coefficient;

[0120] The platform state is determined by the angular velocity amplitude. If the angular velocity is lower than the set amplitude threshold, it enters the static mode; otherwise, it enters the dynamic mode.

[0121] The two-stage filtering structure is combined to enhance the stability of the angular velocity signal and adjust the response performance under different vibration intensities, providing more reliable angular velocity input for the attitude solver module and the error calculation submodule.

[0122] The angular velocity signal is easily affected by high-frequency noise and transient disturbances in a vibrating environment, which affects the stability of the attitude solution. This system adopts a two-stage filtering structure. The front stage is a sliding average filter, which can effectively suppress instantaneous spikes and high-frequency jitter signals; the back stage introduces The filter exhibits excellent dynamic tracking capabilities and state prediction performance. More importantly, the filter's parameters can be adaptively adjusted based on the platform's motion state, allowing the system to automatically switch filter strength between static and dynamic conditions, thus balancing response speed and filter stability. This structure significantly improves the quality of the attitude input angular velocity signal, providing a solid foundation for subsequent dynamic compensation.

[0123] In this embodiment, the flow rate filter includes a zero value processing unit, a small signal holding unit and a valid signal tracking unit;

[0124] The zero value processing unit is used to identify and eliminate invalid water flow signals that are continuously zero. When the continuous zero value exceeds the set continuous threshold, the processing is triggered. The continuous threshold setting range is 30 to 100 sampling points, which is used to eliminate the influence of false data caused by long-term no signal or echo failure;

[0125] The small signal holding unit is used to hold the small signal, with an activation threshold setting range of 0.05m / s to 0.2m / s. When the signal amplitude is detected to be lower than the activation threshold but persists, the small signal validity is automatically maintained;

[0126] The effective signal tracking unit is used to perform dynamic smoothing on the confirmed effective signal. The response speed is controlled by a time constant. The setting range of the time constant is 0.1s~1.0s, which is used to balance the response delay and the data fluctuation suppression capability.

[0127] Water flow signals are highly time-varying and polymorphic, especially in complex external environments. Various states may occur, including sensor echo failure, small disturbances, or effective flow. The flow velocity filter of this system is divided into three strategies: zero-value processing, small signal retention, and effective signal tracking, which can perform graded processing for different flow velocity characteristics. Zero-value processing eliminates invalid data, small signal retention prevents small fluctuations from being misjudged as noise, and effective signal tracking smoothes the actual water flow signal to avoid over-response. Through this type-based adaptive mechanism, the system effectively improves the continuity, authenticity, and anti-interference of flow velocity data, and demonstrates stronger data stability and credibility in actual measurements.

[0128] In this embodiment, the sliding window variance-trend joint detection algorithm specifically includes:

[0129] The continuous water flow measurement data is divided into a sliding window sequence. The minimum number of stable points in each sliding window is set in the range of 10 to 50, and the sliding window step size is set in the range of 1 to 5 data points.

[0130] For the data within each sliding window, the coefficient of variation is calculated as a stability indicator:

[0131] ;

[0132] in, represents the coefficient of variation, represents the standard deviation of the data samples in the sliding window, Represents the average value of the data samples in the sliding window. When , the current sliding window is judged to be a stable interval;

[0133] The trend indicator is calculated using the first-order difference method to determine whether the data has a continuous change trend:

[0134] ;

[0135] in, Represents a trend indicator, represents the total number of sliding window data samples, Indicates the first data samples, Indicates the first data samples, represents a symbolic function, when , the current sliding window is determined to be an interval with no significant trend change.

[0136] The data credibility assessment module dynamically identifies the stability and trend of measurement data from a time series perspective by introducing a sliding window analysis mechanism. The coefficient of variation of the data within the window is calculated to determine the degree of fluctuation, and combined with first-order difference trend detection, it determines whether the current data is stable and trustworthy. The system uses the assessment results as output confidence indicators, which can be used for external output tagging, data selection, or automatic elimination, improving the security and intelligence of back-end data processing. Compared with traditional unfiltered measurement methods, this mechanism empowers the system to identify reliable sections, eliminating the passive collection of measurement results and providing credibility and quality judgment capabilities, significantly enhancing the application value.

[0137] Example 1

[0138] To verify the feasibility of this invention, it was applied to a field test involving dynamic flow velocity monitoring in a water area using an unmanned platform. The test platform consisted of a multi-rotor unmanned aerial vehicle (UAV) equipped with the measurement system described in this invention. This system integrates a multi-source sensor module (including a three-axis MEMS inertial measurement unit, a millimeter-wave radar flow velocity sensor, and an ultrasonic altimeter), a dynamic motion compensation module, an adaptive Kalman filter bank, and a data credibility assessment module.

[0139] The aircraft flew along a pre-set route at an altitude of approximately 3 to 4 meters above the water. During this flight, the system continuously faced challenges with high-frequency vibrations and attitude disturbances due to wind disturbances and platform attitude changes. Without a gimbal, the system employed a Mahony complementary filter and a fourth-order Runge-Kutta integration method within the attitude solver module to perform high-precision attitude calculations and output attitude quaternions for subsequent compensation. During the test, the platform's maximum pitch angle fluctuation reached ±14°, and the maximum yaw angle change rate reached 22° per second, demonstrating the system's practicality under intense dynamic conditions.

[0140] The system also uses an error calculation submodule to compensate for measurement errors caused by the platform's rotation in real time. This module calculates the spatial interaction between the platform's angular velocity and the radar's mounting position, extracting the error velocity component and removing it from the radar's raw velocity data to determine the true water velocity. The maximum instantaneous error velocity reached 0.65 m / s, which was reduced to 0.07 m / s after compensation.

[0141] In order to further improve the stability and accuracy of the signal, the angular velocity filter adopts a two-stage structure: the front-stage sliding average filter suppresses high-frequency jitter, and the rear-stage The filter switches its parameters based on the real-time angular velocity amplitude, enabling adaptive switching between static and dynamic modes. Compared to the raw angular velocity data, the filtered signal's peak fluctuation decreased by 45.3%, the standard deviation decreased by 52.3%, and the jitter amplitude of the attitude solution was reduced by 41.7%.

[0142] Radar velocity data is processed through a velocity filter and classified according to different states. In sections with prolonged echo failures, the system successfully eliminated approximately 120 groups of consecutive zero-value points, accounting for 6.7% of the total sampling points. In low-flow sections (<0.2 m / s), the small signal retention module effectively identified and stably output velocity data, with an average deviation of less than 0.03 m / s. In medium- and high-speed flow areas (1.2-2.0 m / s), the tracking module maintained a response time of less than 0.3 seconds, effectively addressing actual velocity fluctuations.

[0143] A total of 24,000 data sets were collected during the test. After processing by the data credibility assessment module, valid sections with high stability and no significant trend drift were identified. Each data set was scored using a combined coefficient of variation and first-order difference trend analysis method. Comparisons with surface velocity control equipment showed that the system's average error remained below 0.08 m / s.

[0144] Table 1 Statistics of flow rate compensation effect under different platform posture change conditions

[0145] Route number Maximum attitude change (°) Original flow rate error (m / s) Error after compensation (m / s) Flow rate measurement deviation improvement rate A1 ±11.6 0.42 0.09 78.6% A2 ±13.2 0.56 0.08 85.7% A3 ±9.8 0.31 0.05 83.9% A4 ±14.0 0.65 0.07 89.2%

[0146] Table 2 Comparison of angular velocity signal characteristics before and after filter processing

[0147] project Original signal Filtered signal Improvement ratio Peak fluctuation (° / s) 25.8 14.1 45.3% Standard deviation (° / s) 7.3 3.5 52.3%

[0148] Table 3 Data stability evaluation results for different flow rate sections

[0149] Flow rate range (m / s) Valid data ratio Mean CV (%) Mean Q Confidence mean 0.1~0.5 85.4% 3.9 0.14 92.7% 0.5~1.5 80.1% 4.5 0.11 90.4% 1.5~2.5 78.7% 4.7 0.17 88.6%

[0150] According to the experimental data in Tables 1 to 3, the performance of the pan-tilt-less dynamic compensation water flow rate measurement system of this embodiment in a vibration environment has been fully verified, reflecting significant technical advantages and practical application value.

[0151] First, as can be seen in Table 1, during flight missions involving drastic attitude changes, the original velocity measurement error was generally high, reaching as high as 0.65 m / s. However, after compensation by the proposed system, the error dropped significantly to between 0.05 and 0.09 m / s, with the velocity measurement deviation improvement rate exceeding 78%, reaching a maximum of 89.2%. This demonstrates the system's significant effectiveness in dynamic compensation, effectively eliminating the error velocity component introduced by the platform's rotational motion and ensuring the authenticity of the measurement data.

[0152] Secondly, from the comparison of angular velocity filtering effects in Table 2, it can be seen that the sliding average filter and Through filter cascade processing, the peak fluctuation of the angular velocity signal is reduced by 45.3% and the standard deviation is reduced by 52.3%, which significantly suppresses high-frequency vibration interference and maintains the average response delay at the millisecond level. This shows that the system has excellent anti-interference ability and real-time performance at the attitude solution front end, providing a more reliable input data basis for subsequent attitude and error calculations.

[0153] Finally, the data stability analysis in Table 3 further demonstrates the system's broad adaptability to different flow velocity ranges. In both low-velocity areas (0.1-0.5 m / s) and high-velocity areas (1.5-2.5 m / s), the system consistently outputs high-confidence data, with valid data rates generally exceeding 78% and average confidence levels above 88.6%. This demonstrates the system's ability to achieve continuous and reliable data acquisition in complex water surface environments.

[0154] The practical examples fully demonstrate that the system of the present invention can achieve high-precision dynamic compensation for water velocity and maintain the stability and reliability of measurement data in measurement scenarios without a gimbal structure and with significant attitude disturbances, thus possessing broad practical deployment feasibility and engineering application prospects. The system not only maintains highly reliable output despite dramatic attitude changes and significant signal quality fluctuations, but also exhibits excellent real-time performance and algorithmic flexibility. It is suitable for a variety of mobile platforms, including drones and unmanned boats, and has important applications in smart water conservancy, emergency patrols, and field scientific monitoring.

[0155] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A pan-tilt-less dynamic compensation water flow rate measurement system for use in a vibration environment, characterized by: It includes a multi-source sensor module, a dynamic motion compensation module, an adaptive Kalman filter group and a data credibility assessment module; The multi-source sensor module includes a three-axis MEMS inertial measurement unit, a millimeter-wave radar flow velocity sensor, and an ultrasonic altimeter, which are used to respectively collect platform attitude information, water flow original velocity, and platform height; The dynamic motion compensation module includes an attitude solver module and an error calculation submodule. The attitude solver module is used to perform attitude calculation based on the platform attitude information and estimate the attitude change of the platform in a vibrating state in real time. The error calculation submodule removes the error velocity component introduced by the platform's own motion from the original water flow velocity based on the platform attitude information and angular velocity changes, thereby achieving dynamic flow velocity compensation without a pan-tilt structure. The adaptive Kalman filter group includes an angular velocity filter and a flow rate filter. The angular velocity filter dynamically adjusts the filter parameters according to the intensity of the platform's angular motion. The flow rate filter dynamically switches the processing strategy according to the effectiveness of the water flow signal measured by the radar, performing different filtering processes on zero values, small signals, and valid signals respectively. The data credibility assessment module is used to judge the stability of the water flow measurement data after dynamic compensation and filtering based on the sliding window variance-trend joint detection algorithm; The attitude solver module uses the Mahony complementary filter algorithm and the fourth-order Runge-Kutta integration method to solve the platform attitude information. The Mahony complementary filter algorithm is based on the measurement information of the three-axis gyroscope and accelerometer, and realizes angular velocity correction through nonlinear error feedback, including: set up For the platform at all times The attitude quaternion represents the rotation state of the platform from the inertial coordinate system to the body coordinate system. The initial value is the unit quaternion ; Get the angular velocity vector measured by the three-axis gyroscope ,in They are the angular velocities around the X, Y, and Z axes, and the gravity direction vector measured by the accelerometer is obtained ; Based on attitude quaternion Calculate the current estimated gravity direction vector : ； in, Quaternion representing attitude The conjugate quaternion of Represents quaternion multiplication operation; The gravity direction vector measured by the accelerometer and the current estimated gravity direction vector Perform vector cross multiplication to obtain the gravity direction error vector ; Constructing a corrected angular velocity vector with proportional-integral feedback : ； in, represents the proportional gain coefficient, represents the integral gain coefficient, Represents the gravity direction error vector The time integral of , and the corresponding correction angular velocity quaternion is constructed ; The fourth-order Runge-Kutta integration method is used to numerically integrate the quaternion differential equation with an integration time step of , the update process is: ； ； ； ； The final updated attitude quaternion is: ； in, 、 、 and It represents the intermediate increment of Runge-Kutta integral, reflecting the trend of differential change. Represents the updated attitude quaternion; The error calculation submodule includes a radar installation parameter online calibration unit, a beam direction compensation unit and a motion error speed compensation unit; The radar installation parameter online calibration unit is used to correct the installation position parameters of the millimeter wave radar in the body coordinate system in real time based on the updated attitude quaternion output by the attitude solver module. The installation position parameter online calibration function satisfies: ； in, Represents the installation position vector of the millimeter-wave radar in the body coordinate system, and Indicates the nominal installation position of the millimeter-wave radar in the horizontal plane of the aircraft. and Indicates the online correction value of the horizontal installation deviation, Represents column vector form; The beam direction compensation unit is based on the angle between the radar beam direction and the horizontal plane. , calculate the beam direction compensation factor : ； in, Represents the projection of the radar beam direction on the horizontal plane of the platform, which is obtained by rotating the static beam direction vector based on the attitude quaternion. represents the norm; The motion error speed compensation unit calculates the error speed component of the platform's own rotational motion in three-dimensional space to the original water flow speed: ； in, represents the error velocity component, represents the rotation matrix from the body coordinate system to the inertial coordinate system, represents the angular velocity vector; The error velocity component is removed from the original water flow velocity, thereby realizing dynamic compensation for the error introduced into the platform motion without a pan-tilt structure.

2. The pan-tilt-less dynamic compensation water flow rate measurement system for use in a vibration environment according to claim 1, characterized in that: The operating frequency of the millimeter wave radar flow velocity sensor is 24 GHz, the velocity measurement range is 0.1 m / s to 20 m / s, and the measurement accuracy is 0.01 m / s.

3. The pan-tilt-free dynamic compensation water flow rate measurement system for use in a vibration environment according to claim 1, characterized in that: The angular velocity filter adopts a two-stage series structure, including: The first stage is a sliding average filter, which is used to perform preliminary smoothing on the raw angular velocity data to suppress high-frequency noise fluctuations. The width of the sliding average filter window is 5 to 15 data points. The second level is A filter is used to further predict the angular velocity trend based on the sliding average filter result and enhance the response capability to dynamic changes, The filter parameters are dynamically adjusted according to the current motion state of the platform. The adjustment rules are as follows: When the platform is in static mode, set ; When the platform is in dynamic mode, set ; in, represents the position update coefficient, represents the speed update coefficient, Represents the acceleration update coefficient; The platform state is determined by the angular velocity amplitude. If the angular velocity is lower than the set amplitude threshold, it enters static mode; otherwise, it enters dynamic mode.

4. The pan-tilt-less dynamic compensation water flow rate measurement system for use in a vibration environment according to claim 1, characterized in that: The flow rate filter includes a zero value processing unit, a small signal holding unit and a valid signal tracking unit; The zero value processing unit is used to identify and eliminate invalid water flow signals that are continuously zero. When the continuous zero value exceeds the set continuous threshold, the processing is triggered. The continuous threshold setting range is 30 to 100 sampling points, which is used to eliminate the influence of false data caused by long-term no signal or echo failure; The small signal holding unit is used to hold the small signal, with an activation threshold setting range of 0.05m / s to 0.2m / s. When the signal amplitude is detected to be lower than the activation threshold but persists, the small signal validity is automatically maintained; The effective signal tracking unit is used to perform dynamic smoothing on the confirmed effective signal. The response speed is controlled by a time constant. The setting range of the time constant is 0.1s~1.0s, which is used to balance the response delay and the data fluctuation suppression capability.

5. The pan-tilt-less dynamic compensation water flow rate measurement system for use in a vibration environment according to claim 1, characterized in that: The sliding window variance-trend joint detection algorithm specifically includes: The continuous water flow measurement data is divided into a sliding window sequence. The minimum number of stable points in each sliding window is set in the range of 10 to 50, and the sliding window step size is set in the range of 1 to 5 data points. For the data within each sliding window, the coefficient of variation is calculated as a stability indicator: ； in, represents the coefficient of variation, represents the standard deviation of the data samples in the sliding window, represents the average value of the data samples in the sliding window. When , the current sliding window is judged to be a stable interval; The trend indicator is calculated using the first-order difference method to determine whether the data has a continuous change trend: ； in, Represents a trend indicator, represents the total number of sliding window data samples, Indicates the first data samples, Indicates the first data samples, represents a symbolic function, when , the current sliding window is determined to be an interval with no significant trend change.

Images