TL;DR
This paper introduces a novel method for measuring the axial alignment of objects using an autocollimator combined with a Strapdown Inertial Navigation System (SINS). This approach overcomes the traditional limitations of autocollimators, enabling accurate measurements even when the system is not level, which is crucial for dynamic platforms like vehicles or aircraft.
What Problem Does This Paper Address?
Autocollimators are precise optical instruments used for measuring angles, but they typically require a stable, leveled environment. This limitation restricts their use in dynamic scenarios or on platforms where leveling is impractical, such as on ships, large aircraft, or moving vehicles. Existing methods often struggle with non-leveled conditions, leading to significant measurement errors or making measurements impossible. The paper aims to address this by integrating a SINS to provide real-time attitude information, allowing the autocollimator to measure axial angles accurately regardless of the system's orientation.
Methods
The core of the proposed method lies in fusing data from two key components:
- Strapdown Inertial Navigation System (SINS): This system, comprised of gyroscopes and accelerometers, measures the attitude (yaw, pitch, and roll) of the entire measurement setup relative to a geodetic coordinate system. The paper utilizes a customized SINS with fiber-optic gyros, known for their high precision and shock resistance.
- Dual-Axis Photoelectric Autocollimator: This optical instrument measures the angular relationship between its own axis and the axis of a target (a plane mirror in this case). It uses a CMOS sensor to detect the reflected beam from the target.
The measurement model establishes a relationship between the geodetic coordinate system, the SINS coordinate system, and the autocollimator's measurement. By combining the attitude information from the SINS with the autocollimator's measurement of the target's relative angle, the system can compute the target's absolute axial angle with respect to the geodetic frame, even under non-leveled conditions. The paper details the coordinate system transformations and the mathematical model used for this fusion.
Error analysis is performed, considering sources like installation errors, focal length variations, pixel errors in the CMOS sensor, and SINS measurement inaccuracies. A Monte Carlo simulation is employed to evaluate the system's performance under these error conditions.
Datasets
The paper does not explicitly mention or use pre-existing datasets for training. Instead, it presents experimental data collected using a custom-built laboratory setup. This setup includes the SINS, autocollimator, a roll adjustment stage, a plane mirror with a 2D adjustment stage, and a parallel light tube for ground truth measurements. The experimental data consists of 121 sets of measurements, capturing various roll angles and corresponding yaw/pitch adjustments of the target mirror, along with the autocollimator's image plane coordinates.
Key Insights and Ideas
- Bridging the Gap: The primary insight is that by fusing SINS attitude data with autocollimator measurements, the long-standing limitation of autocollimators requiring level surfaces can be overcome. This opens up new possibilities for precise alignment in dynamic and non-ideal environments.
- Dynamic Axis Measurement: The method enables axis measurement on moving bases, which is a significant advancement for applications in aerospace, automotive, and military fields where traditional methods are insufficient.
- Error Mitigation through Fusion: The integration of SINS not only allows for non-leveled measurements but also provides a robust framework for analyzing and potentially mitigating various error sources through data fusion.
- Quantitative Validation: The study provides rigorous validation through both Monte Carlo simulations and laboratory experiments, demonstrating the method's consistency and accuracy.
Key Results
The study reports promising results for the proposed autocollimator-SINS measurement method:
- Simulation Results: Monte Carlo simulations indicated that the majority of measurement errors were within ±0.002°, and the overall measurement accuracy was within ±0.006°.
- Experimental Results: Tests conducted on the laboratory setup, covering a roll angle range of ±10°, showed that most measurement errors were within ±0.004°, with an overall measurement error within ±0.006°.
- Consistency: The experimental results closely matched the simulation outcomes, demonstrating good measurement consistency.
- Error Distribution: The analysis of experimental errors revealed that within a specific roll angle range (±10°), the measurement error was largely independent of the roll angle itself, though it increased with the target's angle change.
Limitations and Future Work
The authors acknowledge several limitations and suggest avenues for future research:
- Experimental Group Size: The relatively small number of experimental measurement groups might not fully capture the system's actual measurement capability.
- Sensor Alignment: Potential misalignment between the CMOS sensor and the autocollimator's optical axis could contribute to discrepancies between simulation and experimental results.
- Time Delay: The paper notes that data fusion and synchronization between the SINS and autocollimator were performed using the measurement model, but potential time delays between sensor readings could impact accuracy, especially on dynamic platforms. Future work should explore advanced data fusion and synchronization techniques.
- Engineering Practicality: While the laboratory validation is successful, further field tests are needed to explore the method's engineering practicality in real-world applications.
- Accuracy Improvement: The authors suggest that measurement accuracy can be further improved through enhanced data fusion and synchronization strategies.
Summary updated