Exponentially narrow delay interval technique to stability and stabilization for nonlinear systems and its application to truck-trailer system

Bin Lü, Yufeng Tian, Yue Yang, Yutong Liu, Xiaojie Su, Linsong Zhang

TL;DR

This paper introduces an "exponentially narrow delay interval" technique that uses a single tunable parameter to partition time-varying delays in nonlinear systems more effectively. By combining this with a new, tighter mathematical inequality, the researchers achieved more stable control and higher delay tolerances in complex systems like truck-trailers.

What Problem Does This Paper Address?

In the world of control engineering, time delays are "stability killers." Whether it's the lag in a sensor reading or the time it takes for a hydraulic actuator to move, delays can cause a system to oscillate wildly or fail entirely. This is especially true for nonlinear systems—those where the output isn't directly proportional to the input (like a truck-trailer backing up).

Engineers often use Takagi-Sugeno (T-S) fuzzy models to simplify these nonlinear systems into a set of local linear rules. However, managing time-varying delays within these models remains a massive headache. Existing "delay partitioning" methods—which chop the delay period into smaller pieces to analyze them—often require too many complex parameters or ignore crucial information about the delay's lower bounds, leading to "conservative" results (meaning the controller is less efficient than it could be).

Methods

The authors propose two primary innovations to tighten the stability criteria:

  1. Exponentially Narrow Delay Interval Technique: Instead of dividing the delay interval $[d_0, d_2]$ into equal parts, they use a recursive formula governed by a single tunable parameter $p \in [0, 1]$. This allows them to create $2^l$ subintervals. By adjusting $p$, they can "zoom in" on specific parts of the delay range, making the optimization process much more flexible and computationally efficient than traditional methods.
  2. Parameter-type Reciprocally Convex Inequality: When analyzing the stability of these systems, researchers use a "Lyapunov-Krasovskii functional" (LKF). Estimating the derivative of this functional usually involves mathematical inequalities that leave "gaps" in precision. The authors introduced a new inequality that uses pre-specified parameters ($v_1, v_2$) to provide tighter bounds and eliminate certain nonlinear constraints that typically make the math harder to solve.

Datasets

As a theoretical control paper, this study does not use a traditional "big data" dataset. Instead, it evaluates its methods using a Truck-Trailer System benchmark.

  • Model: A three-state nonlinear system representing the angular difference between truck and trailer, the trailer's angle, and its vertical position.
  • Comparison: The results were benchmarked against existing literature (specifically Lian et al., 2021) to compare the Maximum Admissible Upper Bounds (MAUBs)—the longest delay a system can handle before becoming unstable.

Key Insights and Ideas

  • The Power of $p$: The single parameter $p$ acts as a "tuning knob." By changing $p$, researchers can artificially shorten or shift the focus of the delay intervals, allowing the controller to utilize more information about the delay's behavior in real-time.
  • Removing Slack Matrices: Traditional improved inequalities often introduce "slack matrices" and extra decision variables that increase the computational load. This new "parameter-type" approach avoids these, simplifying the resulting Linear Matrix Inequalities (LMIs).
  • Global vs. Local Info: Unlike previous methods that only looked at the interval from zero to the current delay $[0, d(t)]$, this technique considers the entire range $[d_0, d_2]$, capturing dynamics that were previously ignored.

Key Results

The effectiveness was measured by finding the largest possible delay ($d_2$) the truck-trailer system could tolerate:

  • Superior Delay Tolerance: Using their method with $p=0.4$, the researchers achieved a Maximum Admissible Upper Bound (MAUB) of 276.7101, significantly outperforming the previous state-of-the-art result of 262.4908.
  • Stability Verification: Simulation plots (state responses) confirmed that even with an oscillating, time-varying delay ($d(t) = 8 + 0.2 \sin(t)$), the truck-trailer system successfully converged to a stable state, whereas older methods might have failed.

Limitations and Future Work

While the paper demonstrates a clear improvement in stability bounds, a few areas remain open for exploration:

  • Parameter Selection: While $p$ is easier to tune than dozens of individual subinterval lengths, the paper doesn't provide an automated way to find the absolute optimal $p$ for any given system; it currently requires some manual testing (as seen in Table 1).
  • Real-World Hardware: The results are currently based on numerical simulations. Future work could involve testing this on physical scale-model truck-trailers or other robotic systems to see how the technique handles unmodeled noise and hardware jitter.
  • Higher-Order Partitions: The study focused on $l=1$ and $l=2$ (2 or 4 subintervals). Exploring the computational trade-offs of much higher-order partitions ($l > 3$) could be beneficial for extremely high-precision applications.

Summary updated

Abstract

<title>Abstract</title> This study addresses the stability and stabilization challenges in nonlinear control systems with time-varying delays via fuzzy model theory. An exponentially narrow delay interval technique is introduced, facilitated by a single tunable parameter $p$, to equivalently transform the extensive delay period $[d_0, d_2]$ into a series of variable narrow subintervals, expressed as: $[d_{0},d_{2}]=\cup_{i=1}^{2^l}[d_{\frac{i-1}{2^{l-1}}},d_{\frac{i}{2^{l-1}}}]$. Diverging from traditional delay interval methods, this narrow delay interval approach simplifies the optimization process by characterizing each subinterval with just $d_0$, $d_2$, and $p$. The core advantage of integrating the narrow interval approach with fuzzy systems is its capability to accurately represent and manage the intrinsic uncertainty in time-varying delays, thus facilitating robust control in real-world scenarios. A parameter-type reciprocally convex inequality is introduced to estimate the derivative of the narrow Lyapunov-Krasovskii functional, enhancing the precision over conventional methods. These innovations contribute to the formulation of new criteria for the stability and stabilization of the T-S fuzzy delayed system. The practicality and effectiveness of the proposed strategies are demonstrated through their application to a truck-trailer system.
BibTeX
@inproceedings{lu2025exponentially,
  author = {Bin Lü and Yufeng Tian and Yue Yang and Yutong Liu and Xiaojie Su and Linsong Zhang},
  title = {{Exponentially narrow delay interval technique to stability and stabilization for nonlinear systems and its application to truck-trailer system}},
  booktitle = {Nonlinear},
  year = {2025},
  doi = {10.1007/s11071-025-11138-8},
}
Figure