Probabilistic Digital Signal Processor for Biomedical Sensor Noise Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mechanical and biomedical sensors face challenges in extracting reliable data due to contaminating signals that often overlap with the signal of interest, and existing filtering techniques are inadequate for handling nonlinear and non-stationary noise, especially in mobile or ambulatory patients, limiting the ability to measure a wide range of biomedical states.
Innovation Solution
A probabilistic model-based system using a dynamic state-space model and probabilistic digital signal processor to filter, estimate, and extract additional information from sensor data, incorporating physical models to process input data iteratively and generate posterior probability distributions for parameter estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional filtering techniques are used to remove noise from sensor data, then some noise reduction is achieved, but the techniques are inadequate for handling nonlinear and non-stationary noise, resulting in loss of useful signal information and inability to extract hidden physiological parameters
Solution Approach 1:
The patent transforms the filtering approach by changing from fixed linear filter parameters to adaptive probabilistic parameters that dynamically adjust to non-stationary noise characteristics. The probabilistic digital signal processor uses time-varying probability distribution functions to model both signal and noise, allowing reliable extraction of physiological parameters even in presence of complex artifacts.
Solution Approach 2:
The patent replaces conventional mechanical filtering approaches (fixed cutoff frequencies, simple smoothing) with a probabilistic computational model. Instead of using deterministic filter equations, the system employs probability distribution functions and Bayesian inference to separate signal from noise, achieving superior performance for nonlinear and non-stationary noise conditions.
2Ease of manufacture
If simple filters are used to remove artifacts, then processing is computationally simple, but artifacts resembling real processes (such as ectopic beats) cannot be removed reliably
Solution Approach 1:
The patent introduces dynamic adaptation into the filtering process through probabilistic models that continuously adjust their parameters based on incoming data statistics. The probability distribution functions evolve over time to track changing signal and noise characteristics, enabling reliable distinction between artifacts and real physiological events without requiring complex manual tuning.
Solution Approach 2:
The probabilistic digital signal processor implements feedback mechanisms where the estimated signal and noise distributions from previous time steps inform the current filtering operation. This recursive Bayesian updating allows the system to learn from past data and improve artifact rejection while preserving genuine physiological events, achieving high accuracy without excessive computational complexity.
3Device complexity
If biomedical monitoring devices measure a limited number of parameters, then device complexity is reduced, but the ability to expand measurement capabilities to additional biomedical state parameters is constrained
Solution Approach 1:
The probabilistic digital signal processor serves as a universal computational engine that can extract multiple different physiological parameters from the same sensor data stream. By using probability distribution functions to model the underlying physiological processes, the system can simultaneously estimate various state parameters (such as heart rate, blood pressure, oxygen saturation) without requiring separate dedicated hardware for each measurement.
Solution Approach 2:
The patent adds a probabilistic dimension to the measurement process, transforming raw sensor signals into a space of probability distribution functions. This additional dimensional representation allows the system to extract multiple physiological parameters from limited sensor inputs by analyzing different characteristics of the probability distributions, effectively expanding measurement capabilities without increasing hardware complexity.
4Adaptability or versatility
If probabilistic model-based processing is used to extract additional information from sensor data, then measurement capability expands to more biomedical states, but computational complexity and processing requirements increase
Solution Approach 1:
The probabilistic processing is segmented into distinct functional modules: signal modeling using probability distribution functions, parameter estimation through Bayesian inference, and artifact rejection via distribution comparison. This modular segmentation allows the complex probabilistic processing to be implemented efficiently using standard digital signal processing techniques and existing computational libraries, reducing overall system complexity.
Data Source
AI summary
A probabilistic digital signal processor is described. Initial probability distribution functions are input to a dynamic state-space model, which operates on state and/or model probability distribution functions to generate a prior probability distribution function, which is input to a probabilistic updater. The probabilistic updater integrates sensor data with the prior to generate a posterior probability distribution function passed (1) to a probabilistic sampler, which estimates one or more parameters using the posterior, which is output or re-sampled in an iterative algorithm or (2) iteratively to the dynamic state-space model. For example, the probabilistic processor operates using a physical model on data from a mechanical system or a medical meter or instrument, such as an electrocardiogram. Output of the physical model yields an enhanced output of the original data, an output to a second physical parameter not output by the medical meter, or a prediction, such as an arrhythmia warning.