Wireless Motion Detection via Neural Network Signal Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion detection systems often suffer from high false-positive rates and inefficiencies in distinguishing between different types of motion, leading to unnecessary activations in security and control systems.
Innovation Solution
A neural network-based system that utilizes wireless signal properties, specifically convolutional neural networks and fully-connected neural networks, to analyze statistical parameters of wireless signals and accurately detect motion by distinguishing between various categories, such as human, animal, or fan movement, using tagged and untagged data for training and detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional motion detection systems use infrared or optical sensors to detect movement, then motion detection capability is provided, but false-positive rates are high and the system cannot distinguish between different types of motion
Solution Approach 1:
The patent transforms wireless signal parameters (CSI amplitude and phase) into statistical features (variance, skewness, kurtosis) that capture motion characteristics. By changing the parameter representation from raw signal to statistical descriptors, the system achieves better discrimination between different motion types and reduces false positives while maintaining detection accuracy.
Solution Approach 2:
The patent introduces machine learning classifiers (SVM, random forest, neural networks) as intermediary components between the wireless signal and motion detection decision. These intermediaries process the statistical features and learn to distinguish between different motion categories, thereby reducing false positives and improving reliability without sacrificing detection capability.
2Extent of automation
If motion detection systems activate security or control systems upon detecting motion, then automated response is provided, but unnecessary activations occur due to false positives
Solution Approach 1:
The patent implements a feedback mechanism where the machine learning classifier continuously learns from labeled training data and refines its motion classification capabilities. The system uses historical data to improve its discrimination between genuine motion events and false positives, thereby reducing unnecessary activations while maintaining automated response capability.
Solution Approach 2:
The patent changes the decision parameters from simple motion detection to multi-class motion classification. By analyzing statistical features of wireless signals and applying machine learning models, the system can distinguish between different motion types (e.g., human motion vs. animal motion vs. inanimate object motion), enabling more reliable automated responses and reducing false activations.
3Adaptability or versatility
If simple motion detection is used without distinguishing motion categories, then system complexity is low, but the system cannot provide intelligent responses to specific types of motion
Solution Approach 1:
The patent segments the motion detection task into distinct processing stages: wireless signal reception, statistical feature extraction, machine learning classification, and motion category identification. This segmentation allows the system to achieve sophisticated motion category discrimination while managing complexity through modular architecture, where each component performs a specific function.
Solution Approach 2:
The patent introduces machine learning models as intermediary components that bridge the gap between simple signal processing and intelligent motion classification. These intermediaries (SVM, random forest, neural networks) add the capability to distinguish motion categories without requiring complex hardware modifications, thereby achieving adaptability while controlling system complexity through software-based intelligence.
Data Source
Figure 1
Figure 2A~2B
Figure 3
AI summary
In a general aspect, motion in a space can be detected based on machine learning of wireless signal properties. In some aspects, sets of tagged neural network input data are obtained at a neural network training system. Each set of tagged neural network input data is based on a statistical analysis of a series of wireless signals transmitted through a space over a respective time period, and each set of the tagged neural network input data includes a tag indicating whether motion occurred in the space over the respective time period. The sets of tagged neural network input data are processed by the neural network training system to parameterize nodes of a neural network system. Parameterizing the nodes configures the neural network system to detect motion based on untagged neural network input data.