Radar Target Behavior Recognition with Time-Frequency LPC Features

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing human behavior identification technologies based on image acquisition and processing are susceptible to light and line of sight issues, leading to low accuracy.

Innovation Solution

A radar system that transmits modulated radar signals at different frequencies, processes radar echo signals to obtain time-frequency domain data, extracts signal attribute and LPC features, and uses a behavior identification model to accurately classify human behaviors using support vector machines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image acquisition and image processing technology is used for human behavior identification, then the system can capture and analyze pedestrian images, but the identification accuracy is low due to being easily affected by light rays and line of sight

Engineering Contradiction:
Improvebehavior identification accuracyVSAvoidlight and line of sight interference
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent replaces the optical image acquisition system with a radar-based electromagnetic wave detection system. The radar system transmits electromagnetic waves that reflect off targets, capturing range, velocity, and acceleration information without being affected by light conditions or line of sight constraints, thereby resolving the vulnerability to environmental interference while maintaining behavior identification capability

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transitions from optical parameters (light intensity, color, visual features) to electromagnetic wave parameters (range, velocity, acceleration, micro-Doppler signatures). By changing the detection parameters from visual to radar-based measurements, the system achieves immunity to light and line of sight issues while obtaining sufficient data for behavior identification

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If radar signals are transmitted and processed to obtain behavior information, then the identification accuracy is improved by minimizing light and weather impact, but the system complexity increases due to signal processing requirements

Engineering Contradiction:
Improvebehavior identification accuracyVSAvoidsignal processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary signal processing operations including Fast Fourier Transform (FFT) to convert time-domain signals to frequency-domain representations, and Short-Time Fourier Transform (STFT) to generate time-frequency domain data. These preliminary transformations organize the raw radar signals into structured formats that facilitate subsequent feature extraction and behavior classification, reducing the complexity of later processing stages

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the complex signal processing task into distinct stages: range velocity acceleration information extraction, time-frequency domain data generation, feature data extraction, and behavior identification. By dividing the processing pipeline into modular segments, each handling a specific aspect of the signal, the overall system complexity is managed while maintaining high identification accuracy

Inventive Principle:
Principle #1Segmentation

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

The radar system provides accurate behavior identification by minimizing the impact of light and weather, enhancing feature data accuracy and classification efficiency.

Implementation Method 1

a transmit antenna of a radar system continuously transmits radar signals

Methodology Applied
Scientific EffectRadar: Radar

Implementation Method 2

After a radar signal is reflected by the target, a receive antenna receives the radar signal

Methodology Applied
Scientific EffectReflection: Reflection

Implementation Method 3

the radar system mixes the received radar echo signal with the radar signal transmitted when the radar echo signal is received, to obtain a beat signal

Methodology Applied
Scientific EffectMixing: Homodyne Detection

Implementation Method 4

M-point fast Fourier transformation (Fast Fourier Transformation, FFT) may be performed on the beat signal after A/D conversion

Methodology Applied
Scientific EffectFast Fourier Transformation:

Implementation Method 5

a short time Fourier transform may be performed on the frequency domain data of the target in one behavior identification period to obtain time-frequency domain data

Methodology Applied
Scientific EffectShort-Time Fourier Transformation:

Data Source

PatentEP3805786B1Target behavior recognition method, apparatus and radar system
Publication Date: 2025.08.13 YINWANG INTELLIGENT TECHNOLOGIES CO LTD
  • EP3805786B1 patent drawingFigure 1
  • EP3805786B1 patent drawingFigure 2~3
  • EP3805786B1 patent drawingFigure 4

AI summary

Embodiments of this application disclose a method and apparatus for identifying behavior of a target, and a radar system, which can be applied to an automated driving scenario. The method includes: receiving a radar echo signal reflected by a target; processing the radar echo signal to obtain time-frequency domain data; processing the time-frequency domain data to obtain signal attribute feature data and linear prediction coefficient LPC feature data, where the signal attribute feature data is used to represent a feature of the radar echo signal attribute, and the LPC feature data is used to represent a feature of the radar echo signal; and inputting the signal attribute feature data and the linear prediction coefficient LPC feature data into a behavior identification model, and outputting behavior information of the target. Based on this application, the accuracy of identifying behavior of a target can be improved.