Radar Signal Processing via Time-Domain Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current signal processing techniques for radar systems, such as FMCW radar, are processor-intensive and memory bandwidth-consuming, limiting their ability to accurately analyze dynamic systems and requiring batched data processing, which results in time-averaged frequency spectra and 'smearing' of signals, making them less effective for estimating parameters like acceleration and jerk.
Innovation Solution
A method that generates predicted waveforms based on physical states of a system, allowing for comparison with received signals in the time domain without spectral analysis, using drive signals near resonant frequencies and transforming waveforms to accurately model and estimate physical states, including position, speed, and acceleration, while being tolerant to modulation inaccuracies and non-linear modulation schemes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spectral analysis (FFT) is used to process radar signals, then position and speed can be determined by looking for peaks in frequency spectra, but the processing becomes processor-intensive and consumes much memory bandwidth
Solution Approach 1:
The patent extracts only the necessary information (position and speed) directly from the time-domain signal through correlation processing, eliminating the need for complete spectral analysis. This extraction approach processes only relevant signal components rather than performing intensive FFT operations on entire signal blocks.
Solution Approach 2:
The patent replaces the mechanical FFT processing system with a correlation-based detection system. Instead of converting signals to frequency domain through intensive mathematical transformations, the system uses template matching and correlation in the time domain to achieve the same measurement objectives with reduced computational load.
2Measurement precision
If FFT algorithm is used, then frequency spectrum can be analyzed, but data must be batched into lengthy sequential chunks resulting in time-averaged measures
Solution Approach 1:
The patent performs preliminary correlation processing on individual signal samples or small batches before full analysis is complete. By pre-processing signals through correlation with expected target signatures, the system prepares data for rapid identification without waiting for complete data batches, reducing the time-averaging effect.
Solution Approach 2:
The patent implements dynamic, real-time correlation processing that adapts to incoming signals continuously rather than processing static batches. The correlation approach allows incremental updating of position and speed estimates as new signal samples arrive, enabling dynamic tracking without the rigid batch-processing constraints of FFT.
3Measurement precision
If spectral analysis is performed, then position and speed can be estimated, but acceleration and jerk cause 'smearing' of the signal between frequency bins
Solution Approach 1:
The patent uses template copying where known target signatures (templates) are correlated with the received signal. Instead of analyzing the signal's frequency content directly, the system compares the received signal against copies of expected target responses, making the detection robust to signal smearing caused by acceleration and jerk.
Solution Approach 2:
The patent introduces correlation processing as an intermediary step between signal reception and target parameter extraction. The correlation operation acts as a mediator that matches signal patterns without requiring precise frequency bin alignment, thereby eliminating the smearing problem that plagues direct spectral analysis.
4Measurement precision
If FMCW radar with frequency modulated signal is used, then distance can be estimated using frequency difference, but the system requires mixing output and input signals and detecting beat signals
Solution Approach 1:
The patent extracts distance information directly from the phase and time characteristics of the reflected signal through correlation processing, rather than requiring frequency mixing and beat signal detection. This extraction method simplifies the receiver architecture by eliminating mixers and beat frequency detection circuits.
Solution Approach 2:
The patent replaces the mechanical signal mixing system with a correlation-based processing system. Instead of using analog mixers to combine transmitted and received signals, the system uses digital or correlation-based methods to directly compare signal phases and extract range information, reducing hardware complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables flexible and accurate analysis of signals, reducing computational and storage requirements, allowing for real-time processing of individual samples, and providing more precise estimation of physical states, including dynamic changes, without the need for batched data processing or frequency domain conversion.
Implementation Method 1
The signal may comprise a periodic component... The drive signal will comprise a periodic component having a peak power close to a resonant frequency of the physical system
Data Source
AI summary
A method of processing a signal representative of at least one physical property of a physical system comprising generating a set of predicted signals, the set of predicted signals comprising at least one member, each member representing a physical state of the physical system, generating a predicted waveform or the signal for each member dependent upon the physical state, and comparing each predicted waveform with the signal to determine the accuracy with which the physical state represented by the member for which the predicted signal was generated matches an actual physical state of the physical system. In an example embodiment, the physical system is a tyre and the state includes the air pressure within the tyre.