Predefined-Time Adaptive Neural Tracking Control for a Single Link Manipulator with an Event-Triggered Mechanism

Yikai Wang, Yuan Sun, Yueyuan Zhang, Jun HuangSensors 2024 · DOI

TL;DR

This paper presents a novel control strategy for robotic manipulators that achieves precise trajectory tracking within a predetermined time frame, even in the presence of system uncertainties. It combines adaptive neural networks with an event-triggered mechanism to significantly reduce communication overhead without sacrificing performance, making it more efficient for real-world applications.

What Problem Does This Paper Address?

Controlling robotic manipulators is challenging due to their inherent nonlinearities, coupling, and uncertainties. Traditional control methods like PID can struggle with complex systems and require extensive tuning. While adaptive control and neural networks have improved robustness, they often guarantee only asymptotic stability, meaning convergence to the desired state might take an unpredictable amount of time. Furthermore, conventional control systems transmit data at fixed intervals, leading to wasted communication resources, especially when the system has already stabilized. This paper addresses these issues by aiming for predefined-time convergence (guaranteeing a specific upper bound for settling time) and reducing communication load through an event-triggered mechanism.

Methods

The core of the proposed solution lies in a combination of several advanced control techniques:

  • Predefined-Time Control: Unlike finite-time or fixed-time control, this approach allows for the direct specification of the maximum settling time by adjusting a controller parameter. This is crucial for applications requiring precise timing.
  • Adaptive Neural Network Control: A Radial Basis Function Neural Network (RBFNN) is employed to approximate the unknown nonlinear dynamics of the robotic manipulator. This allows the controller to adapt to system uncertainties. A novel adaptive law is introduced that simplifies computational complexity by basing the adaptation on the norm of the neural network weights.
  • Event-Triggered Mechanism with Time-Varying Thresholds: To reduce communication load, control signals are only transmitted when a specific condition is met. The paper introduces a time-varying threshold for this mechanism, which is more efficient than traditional fixed thresholds. This mechanism ensures that communication occurs only when necessary, conserving resources without compromising control performance or causing Zeno phenomena (infinite triggers in finite time).
  • Backstepping Technique: This systematic design procedure is used to construct the control laws, ensuring the stability of the closed-loop system by recursively designing virtual controllers.

Datasets

Not discussed in the paper. The study relies on a simulated nonlinear manipulator system with parameters provided in Table 1.

Key Insights and Ideas

  • Precise Timing is Achievable: The ability to predefine the settling time is a significant advantage for applications where exact timing is critical, moving beyond just ensuring stability.
  • Efficiency Through Smarter Communication: The time-varying threshold event-triggered mechanism is a key innovation. It significantly reduces communication overhead compared to fixed-threshold methods and conventional time-triggered approaches, making the system more practical for resource-constrained environments.
  • Simplified Adaptation: The novel adaptive law for the neural network weights reduces computational burden, making the controller more feasible for real-time implementation.
  • Robustness to Uncertainty: The combination of adaptive neural networks and predefined-time control provides robust tracking performance even when the system's exact dynamics are unknown.

Key Results

The simulation results demonstrate the effectiveness of the proposed control scheme:

  • Tracking Error Convergence: The tracking error for the robotic manipulator's position converges to a small vicinity around the origin within the predefined time frame, regardless of the chosen predefined time parameter ($T_c$).
  • Settling Time Control: The settling time of the system is directly influenced by the parameter $T_c$, allowing for explicit control over the convergence speed. A smaller $T_c$ leads to faster convergence.
  • Communication Resource Savings: The event-triggered mechanism, particularly with the time-varying threshold, significantly reduces the number of control signal updates. For $T_c = 3s$, the number of triggers was 986 over 20 seconds, leading to an estimated communication resource saving of over 90.14% compared to traditional high-frequency sampling. This is substantially better than fixed-threshold mechanisms.
  • Bounded Signals: All signals within the closed-loop system, including states, control outputs, and adaptive parameters, remain bounded, confirming the system's stability.

Limitations and Future Work

The authors acknowledge the following limitations and suggest future research directions:

  • Model Imperfections: The study's performance might be compromised in real-world scenarios by factors not fully captured in the model, such as actuator failures or imprecise system modeling.
  • Actuator Dynamics: The current model does not explicitly account for complex actuator dynamics or saturation.
  • State Error Constraints: Future work could explore control strategies that handle specific constraints on state errors.

The authors propose focusing on developing more precise mathematical models and addressing challenges like state error constraints and actuator failures to devise enhanced control strategies for practical robotic manipulator applications.

Summary updated

Abstract

This paper introduces an adaptive trajectory-tracking control method for uncertain nonlinear systems, leveraging a time-varying threshold event-triggered mechanism to achieve predefined-time tracking. Compared to conventional time-triggering approaches, the employment of a time-varying threshold event-triggered mechanism significantly curtails communication resource wastage without compromising the system's performance. Furthermore, a novel adaptive control algorithm with predefined timing is introduced. This method guarantees that tracking errors converge to within a small vicinity of the origin within a predefined timeframe, ensuring all signals in the closed-loop system remain bounded. Moreover, by adjusting a controller-related parameter, we can predefine the upper bound of the convergence time. Finally, the efficacy of the control scheme is corroborated by simulation results obtained from a nonlinear manipulator system.
BibTeX
@inproceedings{wang2024predefined,
  author = {Yikai Wang and Yuan Sun and Yueyuan Zhang and Jun Huang},
  title = {{Predefined-Time Adaptive Neural Tracking Control for a Single Link Manipulator with an Event-Triggered Mechanism}},
  booktitle = {Sensors},
  year = {2024},
  doi = {10.3390/s24144573},
}
Figure