IMU Initialization via Gravity Disturbance Matrix Rotation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inertial navigation systems face challenges in maintaining accurate positioning due to inaccurate estimation of inertial measurement unit (IMU) systematic errors and difficulties in determining heading or azimuth, especially with high correlations between IMU errors and gravity disturbance components, leading to cumulative integration errors in transformation matrices.
Innovation Solution
The method involves rotating the IMU about a specific axis to generate variations in the gravity disturbance matrix, followed by iterative calculations to minimize these variations and estimate an optimal transformation matrix from the accelerometer frame to the inertial frame, addressing errors in roll, pitch, yaw, and misalignment angles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional Kalman filter methods are used for IMU alignment and error estimation, then the system can perform alignment and estimate IMU systematic errors, but the estimation accuracy deteriorates due to high correlations between IMU systematic errors and gravity disturbance components
Solution Approach 1:
The patent performs preliminary identification of IMU systematic errors through rotational maneuvers before normal navigation operations. By executing specific rotation sequences and capturing gravity disturbance data in advance, the system establishes accurate error estimates that remain stable during subsequent navigation, avoiding the correlation problems that plague continuous Kalman filtering during dynamic operations.
Solution Approach 2:
The patent utilizes rotational maneuvers (vibrations) of the IMU platform to generate varied gravity disturbance measurements. By rotating the platform through specific angles and capturing acceleration data during these dynamic maneuvers, the system creates sufficient excitation to independently identify IMU systematic errors from gravity disturbances, breaking the correlation that plagues static alignment methods.
2Ease of operation
If transformation matrix is obtained through integration using quaternion elements, then the transformation from sensor frame to navigation frame can be computed, but large cumulative integration errors accumulate over time
Solution Approach 1:
The patent employs feedback by continuously monitoring the consistency between measured gravity vectors and those computed from the transformation matrix. During rotational maneuvers, the system compares actual accelerometer measurements with expected gravity directions, and uses this feedback to correct the transformation matrix, preventing cumulative integration errors from growing unbounded over time.
Solution Approach 2:
The patent changes the operational parameters by performing rotational maneuvers at specific angles and durations. By varying the rotation angles, speeds, and sequences, the system generates diverse gravity disturbance patterns that enable more accurate transformation matrix determination, reducing integration errors compared to static or routine operational parameters.
3Device complexity
If IMU systematic errors are not accurately determined, then the initialization process is simple, but accurate determination of heading or azimuth becomes difficult
Solution Approach 1:
The patent performs preliminary identification of IMU systematic errors through rotational maneuvers before normal navigation operations. By executing specific rotation sequences and capturing gravity disturbance data in advance, the system establishes accurate error estimates that remain stable during subsequent navigation, avoiding the correlation problems that plague continuous Kalman filtering during dynamic operations.
Data Source
AI summary
A method for initializing an inertial measurement unit (IMU). The IMU is associated with an accelerometer frame. The accelerometer frame corresponds to an inertial frame. The method includes generating variations in rows of a gravity disturbance matrix of the IMU by rotating the IMU about a rotation axis and estimating an optimal transformation matrix from the accelerometer frame to the inertial frame by minimizing the variations of the rows of the gravity disturbance matrix. The rotation axis passes through a center of the accelerometer frame. A column of the gravity disturbance matrix includes a gravity disturbance vector at an initialization moment of an initialization time. The gravity disturbance vector is associated with a gravity vector of Earth.