Ultrasonic Flow Sensor Machine-Learning Detection for Medication Diversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasonic flow sensors for medication administration are sensitive to signal-to-noise ratio issues and hard-coded thresholds, and methods for identifying medication diversion rely on unreliable barcode verification.
Innovation Solution
Implementing a system with a machine learning model to predict start and end times of fluid flow and medication type using time-series data from ultrasonic flow sensors, and verify medication through feature extraction and comparison with stored representations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If shift detection with hard-coded thresholds is used to determine start and end of medication administration, then the method is simple to implement, but the measurement precision deteriorates due to sensitivity to signal-to-noise ratio issues and inability to adapt to different medication types
Solution Approach 1:
The patent transforms the fixed threshold parameter into a dynamic, learnable parameter. Instead of using hard-coded thresholds, the system employs machine learning models that automatically adapt threshold values based on the specific medication type and signal characteristics, resolving the contradiction between implementation simplicity and detection accuracy.
Solution Approach 2:
The patent replaces the mechanical threshold-comparison method with an intelligent machine learning-based detection system. This substitution enables the system to automatically learn optimal detection parameters from data, achieving high precision without sacrificing ease of implementation through automated model training and deployment.
2Ease of operation
If barcode reading is used to identify medication, then the method is simple and quick, but the reliability deteriorates because it cannot verify the match between the label and the actual medication inside the syringe
Solution Approach 1:
The patent introduces an intermediary verification mechanism using ultrasonic flow sensors to measure actual medication delivery. This intermediary measurement acts as a bridge between the barcode identification and the actual medication administration, enabling verification that the medication delivered matches both the label and the prescribed medication type.
Solution Approach 2:
The patent implements a feedback loop where ultrasonic flow measurements provide real-time verification of medication delivery. The system continuously monitors the delivered volume and characteristics, comparing them against expected values, and provides feedback to confirm or reject the medication identification, thereby ensuring reliability while maintaining operational efficiency.
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
Accurately detects fluid flow and identifies medication types, reducing errors and diversion by improving signal processing and verification methods.
Implementation Method 1
receive a time-series generated by an ultrasonic flow sensor corresponding to a flow of at least one fluid through a fluid flow path of the ultrasonic flow sensor
Data Source
AI summary
Systems, methods, and computer program products are provided for detecting fluid flow through ultrasonic flow sensors and/or identifying medication diversion. An example system includes at least one processor configured to: generate, using at least one machine learning model, based on a time-series generated by an ultrasonic flow sensor, a predicted start time associated with a start of a flow, a predicted time period associated with the flow, and/or a predicted end time associated with an end of the flow; and/or identify, based on the time-series, in a database, at least one stored feature representation associated with at least one medication type.