Posture Estimation Using Angular Velocity Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing posture estimation methods for three-dimensional target objects, such as drones or VR applications, suffer from low accuracy due to high computational demands and cumulative errors from integral operations, leading to delayed and unstable estimation results.
Innovation Solution
A posture estimation method that calculates an angular velocity control amount using a proportional-integral-differential control algorithm to correct measured angular velocity values, thereby reducing cumulative errors and improving the accuracy and stability of posture estimation, and obtaining an estimated posture quaternion value without extensive calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If extended Kalman filtering method or gradient descent method is used for posture estimation, then measurement data from sensors is processed, but computational complexity increases leading to estimation delay and reduced estimation frequency
Solution Approach 1:
The patent extracts and corrects only the angular velocity measurement values from the sensor data using acceleration information, rather than processing all sensor data through complex filtering algorithms. By isolating and correcting specifically the angular velocity values that contain cumulative errors, the system achieves accurate posture estimation without the computational burden of full extended Kalman filtering or gradient descent methods.
Solution Approach 2:
The patent changes the parameter being processed from raw acceleration data to corrected angular velocity values. By transforming the problem into correcting angular velocity measurements rather than directly filtering acceleration data, the system reduces computational complexity while maintaining estimation accuracy through the relationship between angular velocity and posture integration.
2Measurement precision
If extended Kalman filtering method or gradient descent method is used for posture estimation, then sensor measurement data is processed, but computation time increases resulting in low estimation frequency
Solution Approach 1:
The patent extracts and corrects only the angular velocity measurement values from the sensor data using acceleration information, rather than processing all sensor data through complex filtering algorithms. By isolating and correcting specifically the angular velocity values that contain cumulative errors, the system achieves accurate posture estimation without the computational burden of full extended Kalman filtering or gradient descent methods.
Solution Approach 2:
The patent performs preliminary correction of angular velocity values before integration to obtain posture information. By pre-correcting the angular velocity measurements using acceleration data and control algorithms, the system eliminates cumulative errors before they propagate through integration, thereby achieving high estimation frequency without sacrificing accuracy.
3Productivity
If complementary filtering method is used for posture estimation, then computational load is reduced, but estimation accuracy decreases
Solution Approach 1:
The patent introduces feedback by using acceleration measurement values to correct angular velocity values iteratively. The corrected angular velocity values are fed back into the integration process, creating a closed-loop system that continuously eliminates cumulative errors. This feedback mechanism enables the system to achieve high estimation frequency with complementary filtering while maintaining or improving accuracy through the correction loop.
4Ease of operation
If integral operation is performed on measured angular velocity to obtain posture information, then posture estimation is achieved, but cumulative errors increase over time
Solution Approach 1:
The patent performs preliminary correction of angular velocity values before integration to obtain posture information. By pre-correcting the angular velocity measurements using acceleration data and control algorithms, the system eliminates cumulative errors before they propagate through integration, thereby achieving high estimation frequency without sacrificing accuracy.
Solution Approach 2:
The patent introduces feedback by using acceleration measurement values to correct angular velocity values iteratively. The corrected angular velocity values are fed back into the integration process, creating a closed-loop system that continuously eliminates cumulative errors. This feedback mechanism enables the system to achieve high estimation frequency with complementary filtering while maintaining or improving accuracy through the correction loop.
Data Source
AI summary
The present disclosure provides a posture estimation method, a posture estimation apparatus, and a computer readable storage medium. Here, the posture estimation method includes: calculating an angular velocity control amount at a current moment based on a measured acceleration value and an estimated acceleration value at the current moment; correcting the measured angular velocity value at the current moment according to the calculated angular velocity control amount to obtain the corrected angular velocity value at the current moment; and obtaining an estimated posture quaternion value at a next moment according to the corrected angular velocity value obtained by calculation and an estimated posture quaternion value at the current moment.