Abstract
Rolling Rocket/Shell (RR/S) systems utilize high-speed rotation to mitigate impact point errors caused by aerodynamic and mass asymmetries. Accurate spatial attitude determination is critical for effective guidance and control. This paper presents a semi-physical simulation framework for an RR/S attitude algorithm using a non-holonomic Inertial Measurement Unit (IMU). By integrating a 902E-1 two-axis turntable with a four-sample rotation vector algorithm implemented in MATLAB/Simulink, the study demonstrates precise real-time attitude tracking under complex rolling and swaying conditions.
Introduction
The stability of unguided rolling rockets and shells is inherently improved by their rapid spin. However, this same rotation complicates the implementation of active guidance systems, which require precise knowledge of pitch, yaw, and roll. Traditional live-fire testing is cost-prohibitive for iterative development, and standard navigation sensors often struggle with the high angular rates and non-holonomic constraints of such projectiles. This research addresses these challenges through a hardware-in-the-loop (HIL) simulation that validates attitude algorithms against physical sensor data in a controlled environment.
Methodology
Hardware-in-the-Loop Setup
The "semi-physical" nature of the simulation is achieved by mounting an XJGZ-3 IMU onto a 902E-1 two-axis turntable. The turntable's chief axis simulates high-speed longitudinal roll (up to 3600°/s), while the pitch axis simulates the swaying motion typical of flight. The IMU utilizes a hybrid sensor approach: a high-range piezoelectric gyroscope for the roll axis and MEMS gyroscopes for the yaw and pitch axes. This configuration allows the system to handle extreme rotation rates while maintaining sensitivity for lateral movements.
Attitude Algorithm
To process the high-frequency data from the IMU, the researchers implemented a four-sample rotation vector algorithm. Unlike simple integration, this method uses multiple sub-samples within a single processing interval to minimize truncation errors during high-speed rotations. The algorithm employs quaternions for attitude representation to avoid "gimbal lock" and ensure continuous tracking.
The simulation environment generates both purely mathematical inputs for theoretical verification and real-time experimental data from the turntable. The turntable’s high-precision encoders provide the "ground truth" necessary to evaluate the algorithm's performance.
Key Insights and Results
The integration of physical hardware allows for the observation of real-world phenomena, such as sensor cross-coupling and stochastic drift, which are often absent in pure mathematical models.
Key findings include:
- High Precision: Mathematical simulations showed an attitude error of approximately 1‰° over a 10-second window, meeting tactical guidance requirements.
- Robust Tracking: Semi-physical tests confirmed the XJGZ-3 IMU's ability to track attitude accurately even at maximum roll rates of 3600°/s.
- Effective Modeling: The system successfully replicated complex "rolling + swaying" dynamics, providing a reliable platform for testing guidance laws without the need for physical launches.
Conclusion and Future Work
The proposed semi-physical simulation provides an exact and real-time attitude information source for RR/S systems. While successful, the study identified limitations regarding sensor cross-coupling and stochastic drift. Future research will focus on developing advanced error compensation models to further refine the accuracy of real-time attitude outputs by filtering hardware-specific noise.
Summary updated