Vehicle Motion Estimation Using Wheel Frequency Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inertial navigation systems suffer from drift errors in gyroscope measurements, leading to inaccurate velocity and position estimates when GNSS signals are unavailable, limiting their reliability to short durations, and require complex and costly calibration procedures.
Innovation Solution
A method and device that processes inertial sensing data through frequency response analysis to extract wheel angular frequency and size, enabling accurate velocity estimation without GNSS integration, using a processing pipeline that includes filtering, frequency domain transformation, and data fusion to enhance accuracy and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional inertial navigation systems use gyroscope measurements for position estimation, then position information can be provided without GNSS signals, but drift errors accumulate over time leading to reduced accuracy
Solution Approach 1:
The patent introduces an intermediary frequency domain analysis process between raw gyroscope measurements and position estimation. By transforming accelerometer data to the frequency domain and using spectral analysis to identify wheel rotation frequencies, the system creates a mediator that extracts reliable motion information without direct integration of drifting gyroscope data.
Solution Approach 2:
The patent replaces the conventional mechanical integration approach (integrating gyroscope angular rates to get attitude, then integrating accelerometer measurements to get position) with a frequency-domain signal processing approach. This substitution uses spectral analysis to directly extract velocity and position information from accelerometer data, avoiding the drift accumulation inherent in traditional mechanical integration methods.
2Measurement precision
If attitude compensation is applied to filter gravity from accelerometer measurements, then velocity integration accuracy improves, but gyroscope drift errors increase over time
Solution Approach 1:
Instead of using attitude compensation to remove gravity from accelerometer data (the conventional approach), the patent inverts the approach by using frequency domain analysis to identify and isolate the gravitational frequency component, then filtering it out. This reverse approach avoids relying on drifting gyroscope-based attitude estimates.
Solution Approach 2:
The patent substitutes the mechanical attitude compensation method (rotating accelerometer data using gyroscope-derived attitude angles) with a frequency-domain filtering method. By transforming data to the frequency domain and applying spectral filtering, the system removes gravity contributions without using drifting gyroscope measurements, thereby maintaining both accuracy and reliability.
3Measurement precision
If complex calibration procedures are performed in the production line to detect sensor errors, then measurement accuracy improves, but manufacturing costs and complexity increase
Solution Approach 1:
The patent enables the navigation system to self-calibrate and self-correct by using frequency domain analysis to automatically identify sensor characteristics and error sources during operation. The system performs its own calibration by analyzing the spectral content of sensor data and adapting its processing accordingly, eliminating the need for external calibration equipment and procedures.
Solution Approach 2:
The patent changes the processing parameters dynamically by adjusting filter characteristics and processing gains based on frequency domain analysis of the sensor data. This allows the system to optimize its performance for different operating conditions without requiring physical recalibration, thereby reducing manufacturing complexity while maintaining measurement precision.
Data Source
AI summary
A system includes inertial sensors and a GPS. The system generates a first estimated vehicle velocity based on motion data and positioning data, generates a second estimated vehicle velocity based on the processed motion data and the first estimated vehicle velocity, and generates fused datasets indicative of position, velocity and attitude of a vehicle based on the processed motion data, the positioning data and the second estimated vehicle velocity. The generating the second estimated vehicle velocity includes: filtering the motion data, transforming the filtered motion data in a frequency domain based on the first estimated vehicle velocity, generating spectral power density signals, generating an estimated wheel angular frequency and an estimated wheel size based on the spectral power density signals, and generating the second estimated vehicle velocity as a function of the estimated wheel angular frequency and the estimated wheel size.