FMCW Radar Gesture Detection Through Dynamic Range Zoning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional radar-based gesture sensing approaches suffer from noise in raw data, overfitting to irrelevant features, high processing costs, and are strongly coupled to radar parameters, with challenges in data augmentation and privacy concerns.

Innovation Solution

An apparatus and method for gesture detection using a FMCW radar sensor that processes radar data to determine a multidimensional array representation, dynamically defines a gesture detection zone based on the range of a person, and searches exclusively within this zone for hand movements, extracting time-series features to reduce processing footprint and focus on relevant data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional radar-based gesture sensing approaches process all radar data, then gesture detection is performed, but processing costs and computational load are high

Engineering Contradiction:
Improvegesture detection capabilityVSAvoidprocessing costs
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent segments the radar data processing by dividing the overall detection space into a gesture detection zone and other zones. Only data within the gesture detection zone is processed for gesture recognition, while data from other zones is discarded. This segmentation significantly reduces computational load and processing costs while maintaining effective gesture detection capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by focusing processing resources exclusively on the gesture detection zone rather than uniformly processing all radar data. The system dynamically adjusts the gesture detection zone based on detected persons, allocating computational resources locally to where gestures are most likely to occur, thereby reducing overall processing costs while maintaining detection effectiveness.

Inventive Principle:
Principle #3Local quality

2Productivity

If conventional radar-based gesture sensing approaches analyze all radar data, then gesture detection is performed, but memory consumption is high

Engineering Contradiction:
Improvegesture detection capabilityVSAvoidmemory consumption
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent segments radar data storage and processing by maintaining separate data structures for the gesture detection zone versus other zones. Only radar data points falling within the dynamically defined gesture detection zone are stored in memory for further processing, significantly reducing memory consumption while preserving all necessary gesture detection information.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and isolates only the relevant portion of radar data that falls within the gesture detection zone, separating it from the rest of the radar data. This extraction approach ensures that memory resources are allocated only to storing and processing the subset of data that is actually useful for gesture detection, reducing overall memory consumption.

Inventive Principle:
Principle #2Taking out (Extraction)

3Productivity

If conventional radar-based gesture sensing approaches process noise in radar raw data, then gesture detection is attempted, but detection precision deteriorates

Engineering Contradiction:
Improvegesture detection capabilityVSAvoidgesture detection precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent extracts and isolates the gesture detection zone from the overall radar detection space, effectively separating relevant gesture data from noisy background data. By processing only the extracted gesture detection zone data and discarding data from other zones, the system improves detection precision while maintaining gesture detection capability.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies local quality by enhancing processing focus on the gesture detection zone where gestures are most likely to occur. The system dynamically adjusts the gesture detection zone based on detected persons, concentrating computational resources and attention on the local region with highest gesture probability, thereby improving detection precision without sacrificing overall detection capability.

Inventive Principle:
Principle #3Local quality

4Device complexity

If conventional radar-based gesture sensing approaches use fixed detection parameters, then system simplicity is maintained, but adaptability to different scenarios deteriorates

Engineering Contradiction:
Improvesystem simplicityVSAvoidscenario adaptability
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamics by making the gesture detection zone adaptive and dynamic rather than fixed. The system continuously updates the gesture detection zone based on detected persons and their positions, allowing the detection parameters to dynamically adjust to different scenarios and users. This dynamic approach enhances adaptability while maintaining relatively simple system architecture through automated zone adjustment.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies self-service by enabling the radar system to automatically adjust its own detection parameters and gesture detection zone based on detected persons. The system autonomously adapts to different scenarios and users without requiring manual reconfiguration, improving versatility while keeping the user interface simple. The radar system serves itself by automatically optimizing detection parameters based on real-time environmental conditions.

Inventive Principle:
Principle #25Self-service

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 reduces processing costs and memory consumption while providing radar parameter-agnostic gesture sensing, enabling data augmentation and synthetic data generation.

Implementation Method 1

a FMCW radar sensor configured to transmit radio frequency waves into a field of view of the FMCW radar sensor and generate the radar data based on received reflections of the transmitted radio frequency waves

Methodology Applied
Scientific EffectRadar: Radar

Data Source

PatentEP4372405B1Apparatus and computer-implemented method for gesture detection, radar system and electronic device
Publication Date: 2025.09.17 INFINEON TECHNOLOGIES AG
  • EP4372405B1 patent drawingFigure 1
  • EP4372405B1 patent drawingFigure 2
  • EP4372405B1 patent drawingFigure 3

AI summary

An apparatus for gesture detection is provided. The apparatus includes interface circuitry configured to receive radar data indicating one or more measurement signal of a Frequency-Modulated Continuous-Wave (FMCW) radar sensor. Further, the apparatus includes processing circuitry coupled to the interface circuitry and being configured to determine a multidimensional array representation of the radar data. The processing circuitry is additionally configured to determine a range of a person nearest to the FMCW radar sensor from the multidimensional array representation of the radar data. In addition, the processing circuitry is configured to define a gesture detection zone based on the determined range of the person. The gesture detection zone covers exclusively ranges up to the determined range of the person. The processing circuitry is configured to search exclusively the gesture detection zone for hand movements of the person. If a hand movement of the person is detected in the gesture detection zone, the processing circuitry is configured to extract a time-series of one or more feature of the hand movement. Further, the processing circuitry is configured to determine a gesture of the person based on the time-series of the one or more feature of the hand movement.