TL;DR
This paper introduces a novel control system for MEMS disk resonant gyroscopes that significantly reduces the start-up time. By employing a two-stage approach for frequency locking and step response, the system achieves an over 80% reduction in start-up time compared to traditional methods, making these high-performance gyroscopes more practical for applications requiring rapid deployment.
What Problem Does This Paper Address?
MEMS disk resonant gyroscopes (DRGs) offer high performance, including a high quality factor (Q-factor), which is crucial for accuracy. However, a high Q-factor also leads to a long start-up time, as the gyroscope takes a considerable amount of time to reach its stable operating frequency and amplitude. This extended start-up period is a significant drawback for applications that require immediate readiness, such as in automotive systems, robotics, or emergency response devices. The paper aims to overcome this limitation by developing a control strategy that accelerates both the frequency locking and the amplitude settling processes.
Methods
The core of the proposed solution lies in a two-stage control system designed to speed up the start-up process:
- Quick Frequency Locking: Traditional phase-locked loops (PLLs) struggle with slow frequency locking due to the low initial amplitude of the driving signal, which results in a small loop gain. To address this, the authors introduce a "coarse-precision mode transition system." In the initial phase, the gyroscope's output signal is converted into a square wave. This square wave, with its larger amplitude and sharper transitions, is fed into the phase detector. This effectively increases the loop gain and the frequency locking range, allowing for much faster initial frequency acquisition. Once the frequency is approximately locked, the system switches back to using the original sinusoidal signal for precise locking and tracking.
- Fast Step Response: The step response time, which refers to how quickly the gyroscope's amplitude settles to its stable operating level, is also a bottleneck. The paper proposes a "large-small mode transition system" for the driving force. In the initial stage of step response, a larger driving force (achieved by increasing the AC amplitude, Vac) is applied to rapidly increase the output amplitude. Once the amplitude reaches a predetermined threshold, the driving force is reduced to a smaller value to achieve quick amplitude stabilization and prevent overshoot or saturation.
The entire system is implemented using digital signal processing on an FPGA, with analog components like amplifiers, ADCs, and DACs.
Datasets
Not discussed in the paper. The paper focuses on the design and experimental validation of a control system using a specific MEMS disk resonant gyroscope.
Key Insights and Ideas
- Square Wave for Coarse Locking: Using a square wave as the input to the phase detector during the initial frequency locking phase is a clever way to artificially boost the loop gain and accelerate locking, especially when the sensor's output amplitude is very low.
- Two-Stage Driving Force for Amplitude Control: The strategy of applying a large driving force initially to quickly build up amplitude and then switching to a smaller force for stabilization is effective in reducing settling time and preventing oscillations.
- Digital Implementation Advantage: The use of an FPGA allows for flexible implementation and seamless switching between different control modes (coarse vs. precise locking, large vs. small driving force), which is crucial for the proposed two-stage approach.
- Trade-off Management: The paper implicitly highlights the trade-off between locking range/speed and stability. The coarse locking phase prioritizes speed and range, while the precise locking phase ensures accuracy. Similarly, the step response balances rapid amplitude increase with stable settling.
Key Results
The experimental results demonstrate a significant improvement in start-up time:
- Frequency Locking: The quick frequency locking system reduced the locking time by 68% compared to a single-step locking mode when the initial frequency deviation was 2 Hz. For a larger deviation of 10 Hz, the reduction was even more substantial, from 90 seconds to 15 seconds.
- Step Response: The fast step response method achieved settling times of less than 4 seconds.
- Overall Start-up Time: The combined quick start-up method reduced the total start-up time by over 80%, bringing it down to less than 9 seconds for a 2 Hz initial frequency deviation. This is a dramatic improvement from the traditional method's start-up time of over 80 seconds.
Limitations and Future Work
- Amplitude Control and Stability: The paper notes that increasing the driving force (Vac) for faster step response can lead to instability in the phase control loop if not managed carefully. The authors suggest reducing the parameter 'E' (related to the square wave amplitude) during the initial stage of step response to maintain loop stability while increasing Vac. This suggests a need for careful tuning and potentially more sophisticated adaptive control strategies.
- Digital Filter Impact: The paper mentions that digital filters are used after demodulation, which can affect signal processing. While the high-frequency terms are filtered, the specifics of filter design and their impact on performance are not deeply explored.
- Generalizability: While the method is demonstrated on a specific MEMS disk resonant gyroscope, its applicability to gyroscopes with different Q-factors or resonant frequencies would require further investigation.
Future work could focus on further optimizing the switching criteria between the coarse and precise modes, exploring more advanced adaptive control algorithms for improved stability and robustness, and investigating the impact of different digital filter designs on the overall performance.
Summary updated