Robust reset control design for piezo-actuated nano-positioner in presence of hysteresis nonlinearity

Basic Information

• Paper ID: 2510.09445
• Title: Robust reset control design for piezo-actuated nano-positioner in presence of hysteresis nonlinearity
• Authors: A. Sebghati, S. H. HosseinNia (Delft University of Technology)
• Classification: eess.SY (Systems and Control), cs.SY (Systems and Control)
• Publication Date: October 10, 2025 (arXiv preprint)
• Paper Link: https://arxiv.org/abs/2510.09445

Abstract

This paper presents a robust nonlinear control design for piezo-actuated nano-positioning systems using frequency-domain analysis methods. Hysteresis nonlinearity in piezoelectric materials degrades tracking accuracy for high-frequency content and reference signals across different stroke ranges. Since inverse model-based hysteresis compensation methods are unreliable when used alone, a control framework with robustness to residual nonlinearity is required. The study demonstrates that robust linear control design faces inevitable performance limitations. The paper establishes a robust control methodology based on complex-order elements to mitigate this limitation, then implements complex-order control using a constant-gain lead-phase (CgLp) reset controller. The control design is based on sinusoidal input describing function (SIDF) and higher-order SIDF (HOSIDF) tools, with constrained optimization problems provided for controller parameter tuning.

Research Background and Motivation

Problem Definition

Piezo-actuated nano-positioners play a critical role in high-precision machinery, requiring accurate tracking of high-speed periodic trajectories with wide amplitude ranges. However, the inherent hysteresis nonlinearity of piezoelectric materials severely impacts system performance:

1. Hysteresis Nonlinearity Issue: The hysteresis characteristics of piezoelectric materials manifest as DC gain uncertainty, with experimental measurements showing DC gain variations up to 215%
2. Limitations of Existing Compensation Methods: Traditional inverse model compensation methods cannot completely eliminate hysteresis effects due to sensitivity to model parameter accuracy
3. Inherent Limitations of Linear Control: Due to the waterbed effect and Bode gain-phase relationships, robust linear control faces inevitable trade-offs in performance improvement

Research Motivation

• Growing demand for wider bandwidth and higher precision in high-precision applications
• Need to develop motion control systems robust to residual hysteresis nonlinearity
• Breaking through the trade-off between robustness and bandwidth in traditional linear control

Core Contributions

1. Theoretical Contribution: Establishes a robust control theoretical framework based on complex-order elements, proving its ability to overcome inherent limitations of linear control
2. Methodological Innovation: Proposes a novel approach to implement complex-order control using CgLp reset controllers
3. Design Tools: Develops frequency-domain design methods based on SIDF and HOSIDF and constrained optimization parameter tuning strategies
4. Performance Enhancement: Achieves approximately 20% bandwidth improvement and significant accuracy enhancement while maintaining robustness

Methodology Details

Task Definition

Design a robust controller for piezo-actuated nano-positioning systems with DC gain uncertainty, where:

• Input: Reference position signal
• Output: Actual position
• Constraints: Phase margin φ ∈ φₘ, φ̄ₘ, DC gain uncertainty κ ∈ 1, κ̄
• Objective: Maximize robust bandwidth while ensuring tracking accuracy and system stability

System Modeling

The piezo-positioner employs an uncertain linear model:

P(jω, κ) = Pₑ(jω) · Pₘ(jω, κ)
Pₘ(jω, κ) = κ · Pₘ(jω)

Where:

• Pₑ(jω) = Kₑ/(jω/ωₑ + 1) represents electrical dynamics
• Pₘ(jω) = Kₘ/((jω/ωₘ)² + 2ξₘ/ωₘ·jω + 1) represents mechanical dynamics
• κ is the uncertain DC gain parameter

Complex-Order Control Theory

Theoretical Foundation

For linear systems with DC gain uncertainty, the paper proposes complex-order control elements:

C̃(jω) = e^(j(1-γ̃)/2 · arctan(ω/ω̃ᵣ))

Where γ̃ ∈ -1,+1 and ω̃ᵣ is the corner frequency.

Theorem 1 Key Conclusion: By adding ideal complex-order elements to a given robust linear control baseline, higher robust bandwidth can always be achieved without compromising open-loop gain characteristics and robustness.

CgLp Reset Controller Implementation

Controller Structure

The CgLp controller consists of a first-order reset element (FORE) and a first-order lead filter in series:

• FORE baseline linear dynamics: Rᵦₗ = 1/(jω/ωᵣ + 1)
• Lead filter: Cₗ(jω) = (jω/αωᵣ + 1)/(jω/ωf + 1)

SIDF Analysis

The sinusoidal input describing function of the CgLp controller is:

C₁(jω) = R₁(jω) · Cₗ(jω)
R₁(jω) = 1/(jω/ωᵣ + 1) · (1 + j·2ω²(1-γ)(1+e^(-π·ω/ωᵣ))/(π(1+γe^(-π·ωᵣ/ω))(ω²+ωᵣ²)))

Parameter Optimization Strategy

A constrained optimization method is employed to adjust CgLp parameters:

max f_c(1)
γ,fᵣ,ff

subject to: ∠L₁(jωc) ∈ [-180° + φₘ, -180° + φ̄ₘ]
           γ ∈ [γₘ, 1], fᵣ > ν·ff, ff < fₘ

Experimental Setup

Experimental Platform

• Equipment: P-621.1CD PIHera single-axis platform with piezoelectric actuator and capacitive sensor
• Stroke Range: -10, 120 μm
• Resolution: 0.2 nm
• Amplifier: E712 digital module, input range -3, 13.5 V, output range -30, 135 V

System Parameters

System parameters identified from experiments:

• Electrical dynamics: Kₑ = 10, fₑ = 935 Hz
• Mechanical dynamics: Kₘ = 0.4986 μm/V, fₘ = 747 Hz, ξₘ = 0.0089
• Uncertainty: κ = 1.6165 (62% DC gain variation)

Performance Metrics

• Bandwidth Metrics: Crossover frequency fc and average crossover frequency f̄c
• Robustness Metrics: Phase margin φ ∈ 60°, 71°
• Accuracy Metrics: Complementary sensitivity function ±1 dB criterion
• Time-Domain Performance: Rise time, overshoot, integral of absolute error (IAE)

Comparison Methods

1. Robust Linear Control: Linear scheme based on second-order PI controller and notch filter
2. Non-Robust Linear Control: Linear controller without considering uncertainty

Experimental Results

Main Results

Frequency-Domain Performance

• Bandwidth Enhancement:
  • Nominal case: fc*(1) increased from 220 Hz to 269 Hz (22.3% improvement)
  • Worst case: fc*(κ) increased from 376 Hz to 441 Hz (17.3% improvement)
  • Average bandwidth improvement approximately 20%
• Robustness Maintenance: Phase margin maintained within 60°, 71° range across the entire uncertainty range

Open-Loop Characteristic Improvement

Through the CgLp controller, open-loop phase near fc*(κ) was successfully enhanced while maintaining nearly unchanged open-loop gain, validating the theoretical expectations of complex-order control.

Time-Domain Verification

Step Response

• Rise Time: Significant improvement compared to robust linear control
• Overshoot: Maintained within 10%-13% range, compared to 10%-18% variation for non-robust linear control
• Robustness: Stable performance maintained across the full DC gain uncertainty range

Sinusoidal Tracking Performance

Using the IAE metric for evaluation, CgLp control demonstrates significant accuracy improvement compared to linear control methods.

HOSIDF Verification

Higher-order harmonic analysis shows:

• |S∞(κ)| ≈ |S₁(κ)|, |T∞(κ)| ≈ |T₁(κ)|
• Validates the effectiveness of SIDF analysis with sufficiently small higher-order harmonic effects

Ablation Studies

Through comparison with different control schemes, the following are verified:

1. Necessity of complex-order elements
2. Effectiveness of CgLp implementation
3. Rationality of parameter optimization strategy

Related Work

Piezoelectric System Control

• Traditional Methods: Inverse model-based hysteresis compensation, but with issues of model complexity and parameter sensitivity
• Linear Control: Methods such as equal-damping control provide robustness through phase lead, but face bandwidth-robustness trade-offs

Reset Control Theory

• Clegg Integrator: Pioneering nonlinear control concept proposed in 1958
• SIDF/HOSIDF Tools: Enable reset control design and analysis in the frequency domain
• CgLp Control: Recent applications in precision motion systems

Contribution Positioning

This paper is the first to combine complex-order control theory with CgLp reset control, providing a solution that overcomes linear control limitations for piezo-actuated nano-positioning systems.

Conclusions and Discussion

Main Conclusions

1. Theoretical Breakthrough: Demonstrates that complex-order control can overcome the inherent robustness-bandwidth trade-off limitations of linear control
2. Practical Method: CgLp reset controller successfully implements the theoretical advantages of complex-order control
3. Performance Enhancement: Achieves 20% bandwidth improvement and significant accuracy enhancement while maintaining robustness

Limitations

1. Experimental Verification: Currently only simulation verification has been performed, lacking actual hardware experiments
2. Noise Consideration: Although noise issues are mentioned, detailed noise robustness analysis is lacking
3. Computational Complexity: Computational costs of HOSIDF analysis and constrained optimization are not discussed in detail

Future Directions

The paper explicitly proposes the following research directions:

1. Investigation of control signal time response and design of additional filters
2. Experimental verification
3. Integration of feedforward control
4. Investigation of disturbance rejection problems

In-Depth Evaluation

Strengths

1. Theoretical Innovation: Complex-order control theory provides new insights for overcoming linear control limitations
2. Method Practicality: CgLp implementation scheme demonstrates good engineering feasibility
3. Analysis Completeness: Frequency-domain analysis combining SIDF and HOSIDF is rigorous and comprehensive
4. Result Convincingness: Simulation results clearly demonstrate method effectiveness

Weaknesses

1. Experimental Limitations: Lacks hardware experimental verification; simulation model accuracy requires validation
2. Parameter Sensitivity: Insufficient analysis of sensitivity to CgLp parameter adjustments
3. Applicability Scope: Applicability to other types of nonlinearity not discussed

Impact

1. Academic Contribution: Provides new theoretical tools and design methods for the precision control field
2. Engineering Value: Significant application value for high-precision manufacturing and nanotechnology
3. Reproducibility: Detailed parameters and design procedures facilitate reproduction

Applicable Scenarios

• Piezo-actuated nano-positioning systems
• Precision motion control systems with DC gain uncertainty
• Tracking control applications requiring high bandwidth and accuracy
• High-precision equipment such as semiconductor manufacturing and atomic force microscopy

References

The paper cites important literature in control theory and precision motion systems, including:

• Clegg (1958): Pioneering work on reset control
• Saikumar et al. (2019, 2021): Theoretical foundations of CgLp control
• Zhang et al. (2024): Recent developments in SIDF/HOSIDF analysis tools

Overall Assessment: This is a high-quality paper with significant theoretical and methodological innovations, providing new solutions to control problems in piezo-actuated nano-positioning systems. Although experimental verification requires further strengthening, its theoretical contributions and methodological innovations possess important academic and engineering value.