Abstract
This study introduces a real-time controller design method under the effects of network time delay and external disturbance. The study first introduces the digital, virtual, intelligent trend of airplane assembly and reveals the status and problems of digital airplane assembly studies. The Cyber-Physical System (CPS) structure is then proposed for digital airplane assembly, and the real-time control issues are discussed. Then, the question of real-time control undertaken by a parallel robot is simplified to a control question with bounded time delay and complex interference, and a mathematical description is presented. Next, a robust $H_\infty$ controller with a disturbance degree of decay $\gamma$ is designed according to the mathematical description. Finally, a simulation is conducted. All of the experiment results show the feasibility of the above proposed methods.
Introduction: The Digital Evolution of Airplane Assembly
In the high-stakes world of aerospace manufacturing, precision is paramount. As airplane assembly shifts from manual labor to highly automated Cyber-Physical Systems (CPS), engineers face a modern challenge: the network itself. Modern airplane assembly is moving toward "digital, virtual, and intelligent" workflows where robots must communicate constantly with sensors and central controllers over a network. However, network imperfections—specifically time delays and mechanical interference—can compromise the stability of high-precision parallel robots.
The Cyber-Physical System Architecture
To address these challenges, the researchers propose a structured CPS framework. This architecture is divided into three distinct layers that manage the flow of information from high-level decision-making down to physical execution.
- Decision-making layer: Where virtual manufacturing models are established and control orders are generated.
- Communication layer: The network backbone (WiFi, 4G, ZigBee, etc.) that transmits state parameters and control orders.
- Sense and control layer: The physical hardware, including sensors (like laser trackers) and actuators (parallel robots).
The Challenge of Network Latency
The primary technical hurdle in this CPS environment is the "stochastic" or random nature of network lag. There is a delay between a sensor sending data and the controller receiving it ($\tau_2$), and another lag between the controller sending a command and the robot executing it ($\tau_1$).
In high-precision environments, these milliseconds of lag can cause a robot to vibrate uncontrollably or diverge from its target. Traditional controllers, such as standard PIDs, often fail to maintain stability when these delays fluctuate.
Robust $H_\infty$ Control Methodology
The researchers focused on a parallel robot—a high-stiffness machine used for positioning large aircraft parts. To solve the instability, they employed Robust $H_\infty$ Control Theory. Key steps in their approach included:
- System Simplification: The complex multi-axis motion was simplified into a Single Input Single Output (SISO) problem by treating the mechanical coupling between the robot's legs as external interference.
- Mathematical Modeling: A state-space model was built that explicitly includes bounded bidirectional time delays.
- Stability Proof: Using Lyapunov functions, the team proved the system remains asymptotically stable under specific conditions.
- LMI Optimization: Stability requirements were translated into Linear Matrix Inequalities (LMIs). Using MATLAB’s LMI toolbox, they calculated an optimal gain matrix ($K$) to balance response speed and stability.
Simulation and Results
The method was evaluated using simulated operational data in MATLAB/SIMULINK, testing the controller against step functions, sinusoidal disturbances ($0.1\sin(t)$), and variable delay profiles ($0.5\sin(t) + 0.5$).
- Failure of Non-Robust Control: Without the $H_\infty$ logic, the system entered a vibrating divergent state, failing to reach its target as it overcorrected for delayed data.
- Success of Robust $H_\infty$: The proposed controller successfully stabilized the robot despite variable delays and noise. The tracking error converged to zero, demonstrating a clear and stable convergence trend.
Conclusion and Future Work
This study demonstrates that $H_\infty$ theory is a powerful tool for CPS-based assembly, as it is designed to minimize the impact of "worst-case" disturbances. While the simulation results are promising, future work will need to address data packet loss in wireless networks and validate the controller on physical hardware to account for real-world factors like friction and thermal expansion.
Paper Citation: Li, H., Zhang, L., Xiao, T., & Dong, J. (2015). Real-time control for CPS of digital airplane assembly with robust H-infinity theory. Tsinghua Science and Technology, 20(4), 376-384.
Summary updated