Neural Monitoring of Mobility Vehicle Subsystems and Battery Health

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Power-driven personal mobility vehicles, such as wheelchairs, often experience electrical and mechanical subsystem failures without warning, leading to loss of mobility and high repair costs, while conventional battery state estimation methods are inaccurate, leaving users stranded.

Innovation Solution

Implementing neural network models with sensors to monitor and predict the state of health of electrical and mechanical subsystems, including batteries, motors, and brakes, using data from voltage, current, temperature, and other sensors to provide proactive maintenance and accurate battery state of charge predictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional battery terminal measurements (voltage) are used to estimate state of charge, then the measurement method is simple, but the accuracy of battery state estimation is poor

Engineering Contradiction:
Improvebattery state of charge estimation accuracyVSAvoidmonitoring system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces simple voltage measurement with a neural network-based monitoring system that processes multiple sensor inputs (voltage, current, temperature) to accurately estimate battery state of charge and health, resolving the contradiction between measurement simplicity and accuracy

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system transitions from monitoring a single parameter (voltage) to monitoring multiple parameters (voltage, current, temperature) simultaneously, using neural networks to process these changing parameters and provide accurate state estimation without excessive complexity

Inventive Principle:
Principle #35Parameter changes

2Reliability

If neural network models with multiple sensors are implemented to monitor subsystem health, then the accuracy of health prediction is improved, but the device complexity increases

Engineering Contradiction:
Improvesubsystem failure prediction accuracyVSAvoidmonitoring system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system performs preliminary monitoring and prediction of subsystem health status using neural networks, enabling proactive maintenance before failures occur, thereby improving reliability while managing complexity through early detection

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The monitoring system uses the vehicle's existing sensors and computational resources to perform self-diagnosis and health assessment, reducing the need for additional complex external monitoring equipment while maintaining high prediction accuracy

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4306350A1Neural network-based monitoring of components and subsystems for power-driven personal mobility vehicles
Publication Date: 2024.01.17 SUNRISE MEDICAL US LLC
  • EP4306350A1 patent drawingFigure 1
  • EP4306350A1 patent drawingFigure 2
  • EP4306350A1 patent drawingFigure 3

AI summary

A neural network model is in communication with a plurality of sensors to evaluate and determining an operating condition and/or life expectancy of one or more electrical or mechanical subsystems of personal mobility vehicle. The neural network model evaluates subsystem operation in the context of a personal mobility vehicle operational state to determine component status and generate an output indicative of one of component state of health or subsystem operation.