Space-Time Neural Network Radar for Gesture Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Integrating radar systems into small or mobile electronic devices is challenging due to power constraints, size limitations, and hardware restrictions, which lead to reduced performance and accuracy in detecting distributed objects and recognizing gestures.
Innovation Solution
A smart-device-based radar system employing a space time neural network that analyzes both magnitude and phase information of radar data using a multi-stage machine-learning architecture, enabling real-time gesture recognition and power conservation by storing feature data in a circular buffer instead of complex radar data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radar operates continuously with high update rate to improve gesture recognition accuracy, then measurement precision is improved, but use of energy increases significantly
Solution Approach 1:
The radar system implements periodic action by operating in alternating active and idle states. During active periods, the radar transmits signals and processes data to detect gestures with high accuracy. During idle periods, the radar remains dormant to conserve power. This periodic operation pattern allows the system to achieve sufficient gesture recognition accuracy while significantly reducing average power consumption compared to continuous operation.
2Measurement precision
If radar processes complex radar data including phase information to improve gesture recognition accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system extracts only the essential features from complex radar data that are necessary for gesture recognition. Instead of processing all available phase and magnitude information, the system identifies and processes only those features that contribute to accurate gesture detection. This extraction approach maintains measurement precision while reducing the computational burden and device complexity.
Solution Approach 2:
The radar data processing is segmented into distinct stages: data acquisition, feature extraction, and gesture recognition. By dividing the complex processing task into smaller, manageable segments, the system can process phase information effectively without overwhelming computational requirements. Each segment handles specific aspects of the data, reducing overall device complexity while maintaining accuracy.
3Measurement precision
If radar uses more antenna elements to improve signal-to-noise ratio and detection accuracy, then measurement precision is improved, but device complexity and size increase
Solution Approach 1:
The radar system merges the functions of multiple antenna elements into a unified signal processing approach. By combining signals from available antenna elements and applying coherent processing techniques, the system achieves improved signal-to-noise ratio without proportionally increasing device complexity. The merging of antenna functions allows effective use of limited antenna resources while maintaining detection accuracy.
Data Source
AI summary
Techniques and apparatuses are described that implement a smart-device-based radar system capable of performing gesture recognition using a space time neural network. The space time neural network employs machine learning to recognize a user's gesture based on complex radar data. The space time neural network is implemented using a multi-stage machine-learning architecture, which enables the radar system to conserve power and recognize the user's gesture in real time (e.g., as the gesture is performed). The space time neural network is also adaptable and can be expanded to recognize multiple types of gestures, such as a swipe gesture and a reach gesture, without significantly increasing size, computational requirements, or latency.