Living Body Estimation Using Compressed Sensing and Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods fail to accurately estimate the number of living bodies and their positions in a target region using radio signals, as reflected waves from living bodies are weak and often buried in noise, and are limited to detecting oscillations at specific frequencies, making it difficult to track living bodies with variable frequencies.
Innovation Solution
The method involves transmitting signals using multiple transmission antenna elements, receiving reflection signals with multiple reception antenna elements, calculating complex transfer function matrices, extracting living-body component matrices affected by vital activities like respiration or heartbeat, and applying compressed sensing using correlation matrices and steering vectors to estimate the number and positions of living bodies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional eigenvalue decomposition methods are used to detect targets from correlation matrices, then the detection process can be performed, but the number of living bodies cannot be accurately estimated because reflected waves are weak and buried in noise
Solution Approach 1:
The patent segments the target region into multiple regions and divides the correlation matrix into multiple sub-matrices, each corresponding to a specific region. This segmentation allows the system to focus on detecting living bodies in each individual region, improving the accuracy of counting living bodies by analyzing regional characteristics separately rather than treating the entire space as a single unit.
Solution Approach 2:
The patent calculates eigenvalues and eigenvectors from the correlation matrix to extract characteristic values that represent living body presence. By transforming the original signal data into eigenvalue-based parameters and comparing these against threshold values, the system can reliably distinguish living body reflections from noise, enabling accurate counting even when reflected waves are weak.
2Measurement precision
If compressed sensing is used to estimate living body positions, then position estimation can be performed, but the method cannot accurately determine the number of living bodies
Solution Approach 1:
The patent combines compressed sensing with regional segmentation by dividing the target space into multiple regions and applying eigenvalue analysis to each region's correlation matrix. This allows the system to simultaneously estimate positions using compressed sensing principles while accurately counting living bodies by identifying which regions contain detected signals above threshold levels.
Solution Approach 2:
The patent merges compressed sensing technology with eigenvalue decomposition methods, integrating the position estimation capabilities of compressed sensing with the living body counting capabilities of eigenvalue analysis. This hybrid approach enables the system to perform both functions simultaneously - estimating positions through signal sparsity exploitation and counting through regional eigenvalue threshold comparison.
3Ease of operation
If conventional methods detect oscillations at specific frequencies, then detection can be performed, but tracking living bodies with variable frequencies becomes difficult
Solution Approach 1:
The patent employs dynamic frequency analysis by continuously monitoring and adapting to the frequency characteristics of reflected waves. Instead of using fixed frequency thresholds, the system analyzes the temporal and spectral variations in the correlation matrix eigenvalues, allowing it to track living bodies regardless of whether their oscillation frequencies remain constant or vary over time.
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 allows for accurate estimation of the number and positions of living bodies by extracting components affected by vital activities from complex transfer functions, enabling robust detection across varying frequencies and improving accuracy compared to single-carrier methods.
Implementation Method 1
transmitting transmission signals to a target region using the M transmission antenna elements; receiving, by the N reception antenna elements, reception signals that include one or more reflection signals resulting from one or more of the transmission signals transmitted from the M transmission antenna elements being reflected by at least one living body
Data Source
AI summary
Complex transfer functions indicating characteristics of propagation between transmission antenna elements and N reception antenna elements are calculated from reception signals received by the N reception antenna elements during a predetermined period. Components affected by vital activity are extracted from the calculated complex transfer functions. A correlation matrix is calculated from changed components affected by vital activity. A steering vector for regions divided from a target region is calculated. A living-body signal intensity vector is estimated by performing compressed sensing for an unknown value that is the living-body signal intensity vector using a correlation matrix vector and an extended steering vector. The number of components constituting the living-body signal intensity vector and having a value of at least a predetermined threshold is estimated to be the number of living bodies, and positions of regions corresponding to the components are estimated to be estimated positions of the living bodies.