10 results about "Complementary filter" patented technology

This invention discloses a posture monitoring method for smart AI glasses based on multimodal fusion, relating to the field of smart wearable technology. The method includes: collecting head motion data, denoising and standardizing it to generate processed head motion data; using a complementary filter algorithm to fuse the processed head motion data with angular velocity variance, and parsing pitch and roll attitude data to generate posture angle data; analyzing the changing trends of the posture angle data through a multi-stage abnormal behavior recognition algorithm, comparing it with historical behavior patterns to obtain abnormal behavior types, and adjusting the recognition criteria to determine whether the user has an undesirable posture; triggering an alarm signal when an abnormal posture is detected; and analyzing the user's behavior patterns and environmental changes when a corresponding feedback signal is triggered to generate a posture monitoring report. This invention improves the monitoring accuracy of smart glasses, provides an efficient feedback mechanism, and enhances user safety.

This invention relates to an attitude calculation method for an unmanned aerial vehicle (UAV) attitude reference system, belonging to the field of UAV navigation. In this invention, the collected sensor data is first filtered to obtain an initial unit quaternion from the navigation coordinate system to the vehicle coordinate system. The accelerometer output is normalized and an error vector is constructed with the normalized gravity vector. The vehicle motion mode is determined by the accelerometer and gyroscope outputs, and the gain coefficient is adaptively adjusted to calculate the gyroscope error correction, thus completing the attitude update and correcting the horizontal attitude. Simultaneously, the magnetometer output is evaluated; if the magnetometer output is valid, the heading angle is corrected, and the corrected three-axis attitude angles are output in real time. This invention uses only quaternions for calculation, avoiding the calculation of rotation matrices and improving calculation efficiency. Compared to traditional complementary filtering algorithms, it incorporates vehicle motion mode judgment and adaptive complementary filter parameter adjustment, avoiding erroneous attitude corrections by the accelerometer under high dynamic conditions. By applying a second-order complementary filter to process accelerometer and magnetometer data for attitude compensation, horizontal attitude correction can be avoided when magnetometer data is unavailable, achieving decoupling of attitude correction.

This invention relates to the field of unmanned aerial vehicle (UAV) technology, specifically to a UAV servo flap system, a UAV wing, and the UAV itself. The system includes a flap actuator, an inertial measurement module, an airspeed detection module, and an onboard control module. The onboard control module fuses inertial data and airspeed data in real time through a data fusion unit, employing a composite filtering architecture combining a Kalman filter and a complementary filter. A dynamic flap angle calculation unit, based on an aerodynamic model and flight state parameters, uses a gradient descent method to optimize and solve for the optimal flap deflection angle. An error compensation unit corrects inertial measurement errors and the effects of structural elastic deformation online, thereby achieving accurate perception and real-time estimation of flight state, improving flap control accuracy and response speed, and enhancing the reliability of the UAV under complex flight conditions.

The application provides a real-time acupoint tracking method and system and a guasha robot, and relates to the technical field of computer vision recognition. The tracking method comprises a static positioning stage, a two-stage training strategy for fusing traditional Chinese medicine meridian prior knowledge is used to improve the YOLO-Pose model, bone landmark point detection accuracy is improved through meridian perception approximate training and adaptive fine adjustment, and a standardized coordinate system is established based on the bone degree and inch method to realize accurate acupoint mapping; in the dynamic tracking stage, a two-stage optimization complementary filter is used to smooth the acupoint coordinates in the video sequence, and high-frequency jitter is inhibited through a speed smoothing mechanism and historical trajectory weighted fusion. The application solves the problems of traditional guasha relying on experience and large dynamic tracking jitter, improves image detection accuracy, optimizes the balance relationship between smoothness, response speed and accuracy of the dynamic trajectory generated by the image, and realizes high-precision and high-real-time acupoint positioning and tracking.

The application relates to a 3D simulation head movement method and device, a storage medium and an electronic equipment, and relates to the technical field of data processing. The method comprises the following steps: acquiring posture data of the actual head posture of a target patient at present through a nine-axis IMU built in a convenient eyepiece; pre-processing and calibrating the posture data respectively to obtain processed data; fusing gyroscope data, accelerometer data and magnetometer data in the processed data through a complementary filtering algorithm to obtain a head posture solution result of the target patient; generating a target quaternion under a world coordinate system according to the head posture solution result, the target quaternion being a quaternion representing the actual head posture; adjusting the posture of a head three-dimensional model of the target patient to the actual head posture according to the target quaternion; and synchronously displaying the head three-dimensional model in the actual head posture and the actual video frame of the target patient at present in a terminal. The application has the effect of improving the accuracy of examination results.

The application provides a magnetic field orientation method and a directional navigation chip with a built-in real-time geomagnetic model. The magnetic field orientation method is applied to the directional navigation chip. The method comprises the following steps: acquiring inertial data output by an inertial sensor and a magnetic field vector output by a magnetic field sensor; acquiring a local real-time geomagnetic reference vector and a magnetic declination angle predicted by a real-time geomagnetic model built in the directional navigation chip; calculating a magnetic field deviation based on the measured magnetic field vector and the local real-time geomagnetic reference vector; judging whether the magnetic field deviation is greater than a deviation threshold; if the magnetic field deviation is not greater than the deviation threshold, calculating an initial attitude angle based on the inertial data and the measured magnetic field vector by using a complementary filtering algorithm; if the magnetic field deviation is greater than the deviation threshold, calculating the initial attitude angle by reducing the weight of the measured magnetic field vector; obtaining a fused attitude angle after attitude fusion calculation based on the initial attitude angle; and calculating a true north azimuth angle based on the fused attitude angle and the magnetic declination angle. The application can directly calculate and output the true north direction by the directional navigation chip.

The application discloses a pose solution method based on complementary filtering and unscented Kalman fusion, and particularly relates to the technical field of waterways, and the specific solution steps are as follows: S1: for a nonlinear system, the system state vector is iteratively updated by using unscented Kalman filtering; the application fuses Mohony complementary filtering and unscented Kalman filtering (UKF) algorithm, replaces the pose angle directly calculated from acceleration and magnetic field intensity information with the pose angle obtained by solving based on the complementary filtering algorithm from the perspective of making the measurement information more accurate, improves the overall pose solution accuracy; meanwhile, the framework of the fusion algorithm is based on the unscented Kalman filtering algorithm, the probability distribution of the nonlinear function is approximated by unscented transformation, the high-order truncation error caused by Taylor expansion of the extended Kalman filtering algorithm is avoided, and the pose solution accuracy for the nonlinear system is improved.

The application discloses a method for obtaining real-time attitude data and real-time angular velocity data of a guided shell, and comprises the following steps: obtaining measured angular velocity data, measured acceleration data and measured magnetic force data of the guided shell; normalizing the measured angular velocity data, the measured acceleration data and the measured magnetic force data to obtain standard angular velocity data, standard acceleration data and standard magnetic force data; designing a finite-time complementary filter based on attitude kinematics of flight of the guided shell; inputting the standard angular velocity data, the standard acceleration data and the standard magnetic force data into the finite-time complementary filter to obtain angular velocity measurement deviation estimation, acceleration estimation and magnetic force estimation; and calculating the real-time attitude data and the real-time angular velocity data of the guided shell based on the angular velocity measurement deviation estimation, the acceleration estimation and the magnetic force estimation. The application can quickly and accurately obtain the real-time attitude data and the real-time angular velocity data of the guided shell.

This invention discloses an automatic vehicle horizontal attitude control method and computer device based on a vehicle MEMS system. The method includes: calculating the initial pitch angle and initial roll angle of the vehicle body to establish an initial horizontal attitude; calculating first observation data for the pitch angle and roll angle; calculating second observation data for the pitch angle and roll angle; fusing the first and second observation data based on a complementary filtering algorithm and optimizing the attitude estimation results using a Kalman filter; determining the target horizontal attitude of the vehicle body according to the driving mode and calculating the target pitch angle; calculating the deviation between the real-time pitch angle and the target pitch angle; if the deviation exceeds a set deviation threshold, generating a suspension control command; sending the suspension control command to the air suspension ECU; controlling the independent raising and lowering of each suspension according to the suspension control command to adjust the vehicle attitude; stopping the adjustment after the vehicle body reaches the target horizontal attitude. This invention enables automated control of the vehicle's horizontal attitude.