Delta-Sigma Sensor Feedback Filtering for Drift-Resilient Readout
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor technologies face challenges in achieving high precision and accuracy due to drift and offset issues, particularly in long-term measurements and miniaturized devices, where environmental factors and manufacturing imperfections lead to inaccurate data and limited dynamic range, hindering their autonomy and energy efficiency.
Innovation Solution
A delta-sigma sensing device with adaptive pulse modulation encoding and distributed encoding is introduced, featuring delta-sigma modulators that attenuate undesired background signals and provide tunable compression of interfering frequency components, increasing dynamic range and energy efficiency, while effectively modeling dependencies between input signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If sensor size is reduced for miniaturization, then device dimensions decrease and mass fabrication is enabled, but signal amplitude decreases and sensitivity to drift and offset increases
Solution Approach 1:
The patent replaces traditional mechanical/electronic drift compensation methods with a digital signal processing approach. The drift compensation term is calculated and subtracted digitally from the sensor output signal, enabling miniaturized sensors to maintain measurement accuracy without complex hardware calibration mechanisms.
2Measurement precision
If calibration is performed frequently to compensate for drift and offset, then measurement accuracy improves, but device autonomy decreases and energy consumption increases
Solution Approach 1:
The patent performs drift compensation in advance by calculating a drift compensation term during initial operation and using it to correct subsequent measurements. This preliminary correction reduces the need for frequent recalibration, thereby conserving energy and maintaining device autonomy while ensuring measurement accuracy.
3Measurement precision
If machine learning algorithms are embedded in the device for drift and offset compensation, then measurement accuracy improves, but device dimensions and energy budget increase substantially
Solution Approach 1:
The patent employs a simplified, computationally inexpensive drift compensation method that does not require embedding complex machine learning algorithms. Instead, it uses a straightforward digital subtraction of a calculated drift term, achieving accurate compensation without increasing device dimensions or energy budget substantially.
4Device complexity
If a threshold is used to differentiate between signal changes and drift, then drift compensation becomes simpler, but effectiveness is reduced when signal amplitude is smaller than drift
Solution Approach 1:
The patent changes the approach from threshold-based signal differentiation to a direct drift term subtraction method. By calculating and subtracting the drift compensation term from the sensor output, the system effectively compensates for drift regardless of signal amplitude, maintaining measurement precision without increasing complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A delta-sigma device that comprises transducers (2) and delta-sigma modulators (3) wherein the transducers (2) comprise: a sensing stage (100) that detect a property of the environment (106) and an interfacing stage (101); and the delta-sigma modulators (3) comprise: an input stage (113, 114), connected to the interfacing stages (101), for receiving an unfiltered intramodulator feedback signal (109) and a low-pass-filtered intramodulator feedback signal (110), and subtracting them from a sum of one or more of the input transduced signals (108) to obtain a delta error signal (111), a feedforward low-pass filtering stage (102); an analog-to-digital encoder (104), that converts the filtered delta error signal into a digital output signal dout(t) (112), an unfiltered intramodulator feedback path, that converts dout(t) (112) into the unfiltered intramodulator feedback signal (109), and a low-pass-filtered intramodulator feedback path, that converts dout(t) (112) into the low-pass-filtered intramodulator feedback signal (110).