Doppler Radar Gesture Recognition With Valid Frame Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Doppler radar gesture recognition systems face high computation loading due to the use of all cells of the Range Doppler Image as input for neural networks, which adversely affects the performance of smart devices by increasing processing demands.
Innovation Solution
The method involves selecting and processing only valid sensing frames through 2D Fast Fourier Transform to generate sensing maps, identifying cells with maximum amplitude, calculating velocity, and labeling frames based on thresholds, reducing input data for the neural network to only valid frames, thereby reducing computation loading.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all cells of the Range Doppler Image are used as input gesture data for the neural network, then the gesture recognition system can process complete gesture information, but the computation loading becomes high
Solution Approach 1:
The patent extracts only the relevant portion of the Range Doppler Image by identifying the region of interest (ROI) where gesture movements occur. Instead of processing all cells in the RDI, the system selectively extracts cells within the ROI boundary, removing unnecessary data from the processing pipeline while retaining essential gesture information for accurate recognition.
Solution Approach 2:
The patent segments the Range Doppler Image into multiple regions, identifying and isolating the region of interest (ROI) that contains gesture-related movements. By dividing the full RDI into manageable segments and processing only the relevant ROI portion, the system reduces computation loading while maintaining gesture recognition accuracy through focused analysis of the segmented gesture-containing region.
2Adaptability or versatility
If the gesture recognition system uses the processor of the smart device, then the smart device can execute gesture-related functions, but the normal operations of the smart device are adversely affected due to high computation loading
Solution Approach 1:
The patent extracts and processes only the essential gesture-related data from the sensing signal, separating gesture recognition processing from general device operations. By extracting only the necessary computing tasks related to gesture interpretation and isolating them from the overall device processing load, the system enables gesture function execution without significantly impacting normal smart device operations.
3Reliability
If the neural network processes all sensing frames, then complete gesture sequences can be recognized, but the computation resources are excessively consumed
Solution Approach 1:
The patent extracts only the essential features and key moments from sensing frames by identifying the region of interest and selecting representative cells within that region. Instead of processing all sensing frames in their entirety, the system extracts critical gesture-related information from selected frames, reducing computation resources while maintaining the ability to recognize complete gesture sequences through selective feature extraction.
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 significantly reduces computation resources needed, minimizing the impact on smart device performance and enhancing the efficiency of gesture recognition systems by discarding invalid frames and focusing on valid gesture inputs.
Implementation Method 1
One of the conventional gesture recognition systems is the Doppler radar gesture recognition system. The Doppler radar gesture recognition system senses motions, shapes, or gestures exercised by a user and generates a corresponding Range Doppler Image (RDI) which comprises a plurality of cells corresponding to the processed reflected signal from the detected objects
Data Source
AI summary
A gesture recognition system executes a gesture recognition method which includes the following steps: receiving a sensing signal; selecting one of the sensing frames from the sensing signal; generating a sensing map by applying 2D FFT to the selected sensing frame; selecting a cell having a largest amplitude in the sensing map; calculating the velocity of the cell and setting the velocity of the selected sensing frame to be the velocity of the cell; labeling the selected sensing frame as a valid sensing frame if the velocity of the selected sensing frame exceeds a threshold value, otherwise labeling the selected sensing frame as an invalid sensing frame; using all of the sensing maps of the valid sensing frames in the sensing signal as the input data for the neural network of the gesture recognition system and accordingly performing gesture recognition and gesture event classification.