FMCW Radar Object Classification via 1D Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current FMCW radar systems face challenges in accurately classifying objects based on their movement and characteristics, particularly in environments with multiple objects and limited processing resources, where distinguishing features from 2D spectrograms to 1D features can lead to decreased information content and classification accuracy.
Innovation Solution
The method involves generating micro-Doppler and micro-range spectrograms for tracked objects, circularly shifting them around a track velocity or mean/median frequency, and extracting one-dimensional feature vectors for classification using a 1D CNN classifier, which is simpler and more resource-efficient than traditional 2D classifiers, enabling high-accuracy object classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D spectrograms are used for object classification, then classification accuracy is improved, but processing complexity and resource consumption increase
Solution Approach 1:
The patent extracts only the essential features from 2D spectrograms by transforming them into 1D feature vectors through frequency analysis and temporal sampling. This extraction process retains the most discriminative information while removing redundant dimensional data, thereby reducing processing complexity while maintaining classification accuracy.
Solution Approach 2:
The patent transforms 2D spectrogram data into 1D feature vectors by projecting the two-dimensional frequency-time representation into a one-dimensional temporal feature sequence. This dimensionality reduction is achieved through systematic sampling and feature aggregation, converting spatial-spectral information into temporal dynamics that can be processed more efficiently.
2Measurement precision
If 2D spectrograms are used for object classification, then classification accuracy is improved, but power consumption increases
Solution Approach 1:
The patent extracts only the essential features from 2D spectrograms by transforming them into 1D feature vectors through frequency analysis and temporal sampling. This extraction process retains the most discriminative information while removing redundant dimensional data, thereby reducing processing complexity while maintaining classification accuracy.
3Measurement precision
If traditional 2D classifiers are used, then classification accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces complex 2D classifiers with simpler 1D temporal feature analysis, using basic signal processing operations that are computationally inexpensive. This substitution uses readily available processing techniques that require less computational power and simpler hardware, reducing both device complexity and cost while maintaining effective classification performance.
4Productivity
If feature extraction from spectrograms is performed, then processing speed is improved, but information content is lost
Solution Approach 1:
The patent extracts only the essential features from 2D spectrograms by transforming them into 1D feature vectors through frequency analysis and temporal sampling. This extraction process retains the most discriminative information while removing redundant dimensional data, thereby reducing processing complexity while maintaining classification accuracy.
Solution Approach 2:
The patent changes the representation parameters of the spectrogram data by transforming frequency-time 2D information into temporal 1D features through systematic parameter extraction. This parameter transformation maintains the essential dynamic characteristics of the target while adapting the data format for more efficient processing.
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 enhances object classification accuracy by compensating for the loss of information from 2D to 1D features, allowing for real-time classification with lower cost and power consumption, even in environments with multiple objects, by utilizing both micro-Doppler and micro-range spectrograms with a 1D CNN classifier.
Implementation Method 1
An FMCW radar transmits an electromagnetic radiation (EMR) signal with a known frequency that is modulated to vary up and down over time. The radar receives a reflected signal corresponding to the transmitted signal
Implementation Method 2
The radar receives a reflected signal corresponding to the transmitted signal and uses the received signal to determine presence, distance, angle of arrival, speed, and direction of movement of objects
Data Source
AI summary
In described examples, a method of operating a frequency modulated continuous wave (FMCW) radar system includes the following steps. A signal is received. A range Fast Fourier Transform (FFT), a Doppler FFT, and an angle FFT are performed on the signal to generate a radar cube. A point cloud is detected corresponding to the radar cube. Multiple objects corresponding to the point cloud are tracked to generate, for respective ones of the tracked objects, a centroid, a boundary, and a track velocity. For respective ones of the tracked object, a micro-Doppler spectrogram and a micro-range spectrogram are generated. The tracked objects are classified based on corresponding micro-Doppler spectrograms and micro-range spectrograms. In some examples, one dimensional feature vectors are extracted from spectrograms and used for training and classification. In some examples, spectrograms are circularly shifted around a track velocity, or a mean or median frequency, prior to feature extraction.