Enhancing Reliability in Redundant Homogeneous Sensor Arrays with Self-X and Multidimensional Mapping

Area of Science:

• * Sensor technology and fault-tolerant systems.
• * Measurement science and data analysis.
• * Engineering and system reliability.

Background:

• * Mechanical defects and sensor failures compromise the reliability of low-cost sensors.
• * Inaccurate measurements can lead to critical system failures, safety hazards, and disruptions.
• * Existing systems lack robust solutions for dynamic error mitigation in redundant sensor arrays.

Purpose of the Study:

• * To present a novel Self-X architecture incorporating sensor redundancy and dynamic calibration.
• * To demonstrate the effectiveness of multidimensional mapping for error mitigation in sensor systems.
• * To validate the proposed approach using synthetic and physical data from tunnel magnetoresistance (TMR) sensors.

Main Methods:

• * Development of a Self-X architecture with dynamic calibration based on multidimensional mapping.
• * Validation using synthetic data from tunnel magnetoresistance (TMR) sensors.
• * Implementation of a physical measurement setup with controlled fault injection for realistic testing.

Main Results:

• * Multidimensional mapping effectively mitigated static and dynamic errors in TMR sensors.
• * Mean Absolute Error (MAE) was reduced by over 80% across sensor combinations.
• * MAE decreased from 1.7°-5.6° for individual sensors to 0.111° using factor analysis with four sensors.

Conclusions:

• * Sensor redundancy and dimensionality reduction algorithms create robust, fault-tolerant measurement systems.
• * The proposed Self-X architecture enhances the reliability of low-cost sensors in critical applications.
• * Dynamic calibration effectively addresses various TMR sensor failure modes, improving long-term system performance.