Seismic Target Tracking via Probabilistic Mixture Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing identity tracking systems fail to effectively distinguish and identify sources in environments with high noise and variable conditions using seismic sensors, as they are sensitive to systematic biases and require prior knowledge of source characteristics.
Innovation Solution
The process employs band-pass filtering, zero-crossing methods for frequency analysis, and probabilistic mixture models to extract and represent seismic data features, forming cumulative probability distributions and using Gaussian mixtures to model emitter characteristics, which are resilient to noise and environmental variations, allowing for real-time identification and tracking without prior knowledge of the emission process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Length of stationary object
If conventional acoustic sensors are used for detection, then the system is simpler and easier to operate, but the detection range is limited and cannot effectively distinguish sources in cluttered environments
Solution Approach 1:
The patent replaces conventional acoustic sensors that detect sound waves through air with seismic sensors that detect vibrations through the earth. This substitution enables detection of targets at greater ranges and through cluttered environments (vegetation, buildings) by utilizing the earth as a propagation medium, which provides stationary and more stable vibration transmission characteristics
2Measurement precision
If seismic sensors are used to detect vibrations through the earth, then detection range and source distinction capability are improved, but the system becomes sensitive to systematic biases and environmental variations
Solution Approach 1:
The patent employs dynamic adaptation through probabilistic mixture models that can adjust to varying environmental conditions. The system uses cumulative probability distributions that evolve with incoming data, allowing it to adapt to systematic biases and environmental variations while maintaining reliable source identification. The dynamic nature of the probability models enables the system to accommodate changes in propagation characteristics without requiring complete re-calibration
Solution Approach 2:
The patent transforms the seismic signal data into cumulative probability distribution functions, changing the parameter representation from raw amplitude/time-domain signals to probability-domain characteristics. This parameter transformation makes the system more robust to environmental variations because probability distributions capture the essential characteristics of sources while being less sensitive to propagation medium variations
3Measurement precision
If high-dimensional data representation is used to maintain resilience against systematic biases, then measurement precision is improved, but computational expense increases
Solution Approach 1:
The patent extracts the essential characteristics of seismic signals by transforming them into cumulative probability distribution functions. This extraction process identifies and retains only the most relevant features (frequency content, temporal patterns) while discarding redundant information. The cumulative probability representation captures the essential source characteristics in a computationally efficient format that maintains resilience against systematic biases without requiring the full complexity of raw high-dimensional data
Data Source
Figure 1
Figure 2~3
Figure 4(a)~4(b)
AI summary
A method of identifying and tracking a target is described, in which seismic data relating to a target is passively detected and processed using statistical means. The statistical manipulation of the data includes frequency information extraction, dynamical mixture model construction based on existing known data and identification of an unknown target by the convergence of this model to a state characteristic of that target.