A Temperature Compensation Approach for Micro-Electro-Mechanical Systems Accelerometer Based on Gated Recurrent Unit–Attention and Robust Local Mean Decomposition–Sample Entropy–Time-Frequency Peak Filtering

Rubiao Cui,Jingzehua Xu,Botao Huang,Huakun Xu,Miao Peng,Jingwen Yang,Jintao Zhang,Yikuan Gu,Daoyi Chen,Haoran Li,Huiliang Cao

DOI: https://doi.org/10.3390/mi15040483

IF: 3.4

2024-03-31

Micromachines

Abstract:MEMS accelerometers are significantly impacted by temperature and noise, leading to a considerable compromise in their accuracy. In response to this challenge, we propose a parallel denoising and temperature compensation fusion algorithm for MEMS accelerometers based on RLMD-SE-TFPF and GRU-attention. Firstly, we utilize robust local mean decomposition (RLMD) to decompose the output signal of the accelerometer into a series of product function (PF) signals and a residual signal. Secondly, we employ sample entropy (SE) to classify the decomposed signals, categorizing them into noise segments, mixed segments, and temperature drift segments. Next, we utilize the time-frequency peak filtering (TFPF) algorithm with varying window lengths to separately denoise the noise and mixed signal segments, enabling subsequent signal reconstruction and training. Considering the strong inertia of the temperature signal, we innovatively introduce the accelerometer's output time series as the model input when training the temperature compensation model. We incorporate gated recurrent unit (GRU) and attention modules, proposing a novel GRU-MLP-attention model (GMAN) architecture. Simulation experiments demonstrate the effectiveness of our proposed fusion algorithm. After processing the accelerometer output signal through the RLMD-SE-TFPF denoising algorithm and the GMAN temperature drift compensation model, the acceleration random walk is reduced by 96.11%, with values of 0.23032 g/h/Hz for the original accelerometer output signal and 0.00895695 g/h/Hz for the processed signal.

chemistry, analytical,nanoscience & nanotechnology,instruments & instrumentation,physics, applied

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the decline in the measurement accuracy of micro - electro - mechanical systems (MEMS) accelerometers under the influence of temperature changes and noise. Specifically, in practical applications, MEMS accelerometers are affected by temperature fluctuations and internal and external noise, resulting in a significant deviation between their output signals and ideal signals, thereby reducing the measurement accuracy. To address this challenge, the paper proposes a parallel denoising and temperature - compensation fusion algorithm based on the gated recurrent unit (GRU) attention mechanism, robust local mean decomposition (RLMD), sample entropy (SE), and time - frequency peak filtering (TFPF).

### Overview of the Solution

1. **Signal Decomposition**:
- Use RLMD to decompose the output signal of the accelerometer into a series of product function (PF) signals and a residual signal.

2. **Signal Classification**:
- Utilize sample entropy (SE) to classify the decomposed signals into noise segments, mixed segments, and temperature - drift segments.

3. **Denoising Treatment**:
- Apply the time - frequency peak filtering (TFPF) algorithm with different window lengths to the noise segments and mixed segments for denoising treatment respectively.
- Since the temperature - drift segment is a smooth curve, no denoising treatment is carried out.

4. **Temperature Compensation**:
- Introduce the time series of the accelerometer output as the input of the temperature - compensation model, and combine the gated recurrent unit (GRU) and the attention module to propose a new GRU - MLP - Attention Model (GMAN) architecture.
- Predict the temperature drift through the GMAN model and perform temperature compensation on the denoised signal.

### Experimental Results

The effectiveness of the proposed fusion algorithm has been verified through simulation experiments. After being processed by the RLMD - SE - TFPF denoising algorithm and the GMAN temperature - drift compensation model, the acceleration random walk (representing the noise characteristic) has been reduced by 96.11%. The value of the original accelerometer output signal is 0.23032 g/h/√Hz, and the value of the processed signal is 0.00895695 g/h/√Hz.

### Main Contributions

1. **Innovative Models and Algorithms**:
- Propose a new GMAN model and RLMD - SE - TFPF fusion algorithm. For the first time, use time - series information in the temperature compensation of MEMS accelerometers to predict temperature drift, and the experimental results verify the superiority of this method.

2. **Parallel Processing**:
- The complete temperature - compensation model adopts a parallel - processing method, separates the noise and drift components of the signal, and simultaneously performs denoising and compensation, avoiding additional loss of the original signal and without a delay in the temperature - compensation process.

3. **Performance Improvement**:
- Verify the effectiveness of the method through the Allan algorithm and significantly improve the measurement accuracy of the accelerometer.

In conclusion, by proposing an innovative parallel - denoising and temperature - compensation fusion algorithm, this paper effectively solves the problem of the decline in the measurement accuracy of MEMS accelerometers under the influence of temperature changes and noise, providing new ideas and methods for improving the performance of MEMS accelerometers.