Abstract
Most approaches for self-triggered control (STC) of nonlinear networked control systems (NCS) require measurements of the full system state to determine transmission times. However, for most control systems only a lower dimensional output is available. To bridge this gap, we present in this paper an output-feedback STC approach for nonlinear NCS. An asymptotically stable observer is used to reconstruct the plant state and transmission times are determined based on the observer state. The approach employs hybrid Lyapunov functions and a dynamic variable to encode past state information and to maximize the time between transmissions. It is non-conservative in the sense that the assumptions on plant and controller are the same as for dynamic STC based on hybrid Lyapunov functions with full state measurements and any asymptotically stabilizing observer can be used. We conclude that the proposed STC approach guarantees asymptotic stability of the origin for the closed-loop system.
Introduction
In Networked Control Systems (NCS), controllers and sensors communicate over digital networks with limited bandwidth. To optimize resource usage, researchers often employ Event-Triggered Control (ETC), where updates are sent only when a specific error threshold is crossed. However, ETC requires continuous monitoring of the system state. Self-Triggered Control (STC) offers a more efficient alternative: at each update, the controller proactively calculates the next transmission time, allowing communication hardware to remain idle in the interim.
A significant hurdle in applying STC to complex nonlinear systems is the assumption of full-state feedback. In practical applications, controllers often only have access to limited sensor data (output feedback). This paper bridges the gap between theoretical full-state STC and practical output-feedback scenarios.
Methodology
The proposed framework integrates three core components to achieve efficient control without full state measurements:
- State Observer: A continuous-time observer reconstructs the plant's internal state from available output data. The design is modular; any asymptotically stable observer can be integrated into the framework.
- Hybrid Lyapunov Functions: The system is modeled as a hybrid system with continuous "flows" (physical dynamics) and discrete "jumps" (network transmissions). Lyapunov functions are utilized to mathematically guarantee that the system remains stable and converges to its equilibrium.
- Dynamic Variable ($\eta$): This variable acts as a memory mechanism, encoding a discounted average of past performance. If the system has historically maintained low error, the STC mechanism can "borrow" from this history to extend the interval between transmissions.
The core algorithm computes the next transmission time by evaluating multiple stability parameters to maximize the idle time without compromising system integrity.
Key Results and Insights
The approach was evaluated using a numerical simulation of a single-link robot arm. The results demonstrate a significant reduction in network load compared to traditional periodic sampling.
- Communication Efficiency: In a 50-second simulation, the proposed STC required an average of only 92.4 transmissions, whereas periodic sampling required 282 transmissions to achieve comparable stability. This represents a 3x reduction in network load.
- Increasing Intervals: As the system converges and the observer error decreases, the STC mechanism successfully identifies opportunities to increase the time between transmissions.
- Stability Guarantee: The paper provides a rigorous proof of Uniform Global Asymptotic Stability (UGAS), ensuring the system returns to its equilibrium point regardless of initial conditions.
Conclusion and Future Work
This research demonstrates that moving from full-state feedback to output feedback does not require stricter assumptions on the plant or controller, making the approach highly versatile. While effective, the authors note that large initial estimation errors in the observer can lead to frequent early updates. Future work will focus on mitigating these "initial spikes" and extending the framework to handle network imperfections such as delays, jitter, and packet loss.
Summary updated