---
title: MIMO Dual Iterative Learning Control
url: https://www.emergentmind.com/topics/mimo-dual-iterative-learning-control-dilc
type: topic
---

# MIMO Dual Iterative Learning Control

Searching arXiv for the cited DILC and related MIMO ILC papers to ground the article in current literature.
MIMO Dual Iterative Learning Control (DILC) is a data-driven iterative learning scheme for repetitive Multiple-Input Multiple-Output (MIMO) systems that simultaneously learns the tracking control input and the plant model, without requiring any prior system knowledge or manual parameter tuning. In the formulation reported in "Dual Iterative Learning Control for Multiple-Input Multiple-Output Dynamics with Validation in Robotic Systems" [2509.18723], the method is designed for repetitive MIMO systems, integrates with established iterative learning control methods, and provides monotonic convergence conditions for both reference tracking error and model error in linear time-invariant systems. Within the broader MIMO ILC literature, MIMO DILC belongs to a family of trial-domain learning methods concerned with interaction, uncertainty, scalability, and convergence; related work includes robust convergence analysis for nonsquare MIMO ILC via system equivalence transformation [1910.10305], multivariable ILC design procedures ranging from decentralized to centralized synthesis [1806.08550], and data-driven stochastic-gradient MIMO ILC methods [2111.08445].

## 1. Definition and problem setting

MIMO DILC addresses the problem of autonomously solving repetitive motion tasks on unknown MIMO dynamics. The underlying setting in [2509.18723] is a repetitive, unknown MIMO system with inputs \(u_j(n) \in \mathbb{R}^O\), outputs \(y_j(n) \in \mathbb{R}^O\), trial index \(j\), sample index \(n\), and \(N\) samples per trial. Under the linear time-invariant assumption, the dynamics are written as \(y_j = P u_j\), where the plant matrix \(P\) is unknown and must be learned. The objective is to find input trajectories \(u_j\) such that \(y_j\) tracks a reference trajectory \(r\), while avoiding prior model identification and manual tuning [2509.18723].

The adjective “dual” refers to the coupling of two learning processes: model learning and control learning. Rather than treating system identification and trial-domain control refinement as separate procedures, DILC intertwines them on every trial. This distinguishes it from single-loop ILC formulations that update only the feedforward input. In the language of [2509.18723], the framework is simultaneously a tracking controller and a model learner.

The MIMO setting is central rather than incidental. Multivariable interaction can undermine stability and performance if it is ignored, and the broader MIMO ILC literature has emphasized a spectrum of design procedures that trade modeling effort against achievable performance [1806.08550]. DILC is positioned at the data-driven end of that spectrum: it is intended to cope with unknown dynamics and to eliminate the need for manual parameter tuning, including in complex robotic systems [2509.18723].

## 2. Dual learning architecture and lifted representation

The DILC algorithm in [2509.18723] is organized around three steps per trial: model learning via Iterative Model Learning (IML), control input learning via model-based ILC, and self-parametrization.

In the model-learning stage, two lifting operators are introduced: one arranges the model as a large parameter vector, and the other constructs a Toeplitz matrix for input, exploiting system structure. The unknown plant matrix \(P\) is reshaped into a vector \(p\), and the model estimate \(M_j\) into \(m_j\). The model prediction is written as
\[
\hat{y}_j(u) = M_j u \approx U m_j,
\]
where \(U\) is built from \(u\) via the lifting operation [2509.18723].

The prediction error is
\[
\epsilon_j(y,u) = y - U m_j,
\]
and the model update law is
\[
m_{j+1} = m_j + K_j \epsilon_j(y_j, u_j).
\]
The learning gain \(K_j\) is generated automatically using standard ILC design functions adapted to the model-learning context. For gradient IML,
\[
K_j = U_j^\top W_j,
\]
and for norm-optimal IML,
\[
K_j = (U_j^\top W_j U_j + S_j)^{-1} U_j^\top W_j,
\]
with \(W_j\) and \(S_j\) positive definite weights [2509.18723].

In the control-learning stage, the latest model estimate \(M_{j+1}\) is used to construct the input update. With tracking error
\[
e_j = r - y_j,
\]
the update law is
\[
u_{j+1} = u_j + L_j e_j.
\]
For gradient ILC,
\[
L_j = M_{j+1}^\top W_j,
\]
and for norm-optimal ILC,
\[
L_j = (M_{j+1}^\top W_j M_{j+1} + S_j)^{-1} M_{j+1}^\top W_j.
\]
The paper describes these gains as self-parametrized: all gain matrices are computed automatically, based on current model sensitivity, to ensure fast and stable convergence and to adapt to scale and coupling variations typical in MIMO systems [2509.18723].

Algorithmically, the procedure is trial-by-trial. One initializes \(u_0\) and \(m_0\), applies \(u_j\), measures \(y_j\), updates \(m_{j+1}\), computes \(M_{j+1}\), determines \(L_j\), and updates \(u_{j+1}\). The paper characterizes this as plug-and-play and purely data-driven [2509.18723].

## 3. Convergence properties and theoretical guarantees

The theoretical analysis in [2509.18723] separates model convergence, prediction error convergence, and tracking error convergence.

For model learning, the stated goal is monotonic convergence of the model error \(p - m_j\). Under sufficient excitation and appropriately designed gains, the main condition is
\[
\| I - K_j U_j \| \leq 1,
\]
together with \(K_j\) full column rank. Under these conditions,
\[
\| p - m_{j+1} \| \leq \| p - m_j \|,
\]
with strict decrease after enough trials [2509.18723].

For prediction, the requirement is exponential contraction of the prediction error:
\[
\|\epsilon_{j+1}(y, u)\| \leq \gamma \|\epsilon_j(y, u)\|,\quad \gamma < 1,
\]
provided
\[
\|I - U_j K_j\| \leq \gamma.
\]
This gives a direct statement that the model-induced output prediction becomes progressively more accurate over trials [2509.18723].

For tracking, the control-learning stage is said to achieve exponential monotonic tracking convergence:
\[
\| e_{j+1} \| \leq \alpha \| e_j \|,\quad \alpha < 1.
\]
The key statement is that once the model estimate enters a bounded neighborhood of the true plant, the control update ensures monotonic tracking convergence. The paper further states that this is independent of initial model quality and requires no tuning [2509.18723].

An important technical qualification is that these monotonic convergence conditions are provided for linear time-invariant systems. The same paper validates the method on nonlinear high-fidelity simulation and multiple nonlinear real-world MIMO systems. This suggests that the empirical operating regime extends beyond the formal LTI analysis, although that extension is experimental rather than a stated theorem [2509.18723].

## 4. Relation to MIMO ILC theory, interaction, and nonsquare systems

MIMO DILC is naturally read against earlier MIMO ILC theory on interaction and convergence. In multivariable ILC, one standard lifted update law is
\[
f_{j+1} = Q \left( f_j + L e_j \right),
\]
with robust convergence characterized by
\[
\rho(Q(e^{j\omega})(I - L(e^{j\omega})J(e^{j\omega}))) < 1 \quad \forall \omega,
\]
and monotonic convergence by
\[
\gamma := \| Q(I - L J) \|_\infty < 1.
\]
The work in [1806.08550] develops a range of procedures—from independent SISO ILC to robust multi-loop SISO, decentralized robust MIMO ILC using Gershgorin-type bounds and the structured singular value, and centralized MIMO ILC—precisely to address interaction without or with increasing degrees of MIMO model knowledge.

A more specific theoretical issue arises in general MIMO ILC convergence analysis for nonsquare systems. "System Equivalence Transformation: Robust Convergence of Iterative Learning Control with Nonrepetitive Uncertainties" [1910.10305] identifies a contradiction between convergence conditions for the output tracking error and for the input signal or input error. In the summary provided for that paper, indirect input-based analysis typically requires full column rank, whereas direct output-error-based analysis typically requires full row rank; for nonsquare MIMO systems, both cannot hold unless the system is square and invertible. The proposed system equivalence transformation (SET) transforms a general nonsquare MIMO tracking problem into an equivalent square MIMO one, updates only the minimum required \(p\) input channels for a \(p\)-channel reference, and yields a unified convergence condition for boundedness of all signals and robust convergence of the tracking error [1910.10305].

A plausible implication is that DILC design and convergence analysis can use SET-type reasoning when nonsquare MIMO structure or nonrepetitive uncertainties create contradictions between input-domain and output-domain conditions. The summary for [1910.10305] explicitly notes that DILC design and convergence analysis can use the SET technique for easy stability checks in the transformed domain. This does not make SET part of the DILC algorithm in [2509.18723]; rather, it situates DILC within a broader theoretical apparatus for MIMO learning control.

## 5. Experimental validation in robotic and nonlinear systems

The experimental validation reported in [2509.18723] spans both simulation and real-world systems.

In a 6-DOF industrial robot simulation (UR10e, MuJoCo), the dynamics are described as highly nonlinear and strongly coupled, with aggressive, discontinuous references. The reported outcome is convergence to small tracking error within 10–100 trials without model knowledge or tuning, while tracking all six joint references well [2509.18723].

In a two-link real-world robot with friction and backlash, three different references are tested, and all converge within 50 trials. In a three-wheeled inverted pendulum robot, the experiments involve balancing, non-repetitive initial states, and strong coupling; for three complex reference tasks, the reported reductions are more than 50% error reduction in 10 trials and more than 80% in 20 trials [2509.18723].

Across all systems and reference trajectories, the paper states that validation is autonomous, requires no tuning and no task-specific parameters, applies to both gradient-based and norm-optimal update policies, and is robust to nonlinearities, time-variance, and modeling errors, provided the system is repetitive. The abstract also states that many reference tracking tasks are solved within 10–20 trials, and even complex motions are learned in less than 100 iterations [2509.18723].

These experimental results are significant because the theoretical development is framed in terms of LTI systems, whereas the validation includes nonlinear simulation and nonlinear real-world MIMO systems. The intended interpretation is not that the LTI proofs directly cover all tested scenarios, but that the dual learning mechanism and self-parametrization are empirically effective in settings with coupling, measurement noise, and plant complexity [2509.18723].

## 6. Distinctions from related learning-control frameworks

MIMO DILC is adjacent to, but distinct from, several other families of iterative learning methods.

First, it differs from model-free stochastic-gradient MIMO ILC methods whose primary object is feedforward optimization rather than simultaneous model learning. In "Conjugate gradient MIMO iterative learning control using data-driven stochastic gradients" [2111.08445], the cost is written as
\[
\mathcal{J}(f) = \|r - J f\|_2^2,
\]
with true gradient
\[
g(f) = 2J^\top J f - 2J^\top r = -2 J^\top e,
\]
and an unbiased stochastic gradient estimate constructed from randomized experiments. The method develops conjugate search directions and optimal step-size selection using dedicated experiments. That paper explicitly does not mention or discuss Dual Iterative Learning Control specifically [2111.08445]. The contrast is useful: DILC unites control learning and model learning, whereas stochastic conjugate-gradient ILC addresses data-driven input learning with experimental efficiency.

Second, MIMO DILC should not be conflated with two-loop adaptive ILC frameworks in which “dual” refers to coupled physical subsystems or coordinated learning loops rather than simultaneous model and control learning. The segment-based two-loop adaptive ILC framework for spacecraft position and attitude tracking combines position and attitude errors into a dual-number representation, employs two learning loops that interact through a dual-number representation of tracking errors, and uses a segment-based dynamic projection mechanism to keep parameter estimates and control inputs bounded [2602.14660]. That framework targets unknown but repeatable parameters and disturbances in strongly coupled rigid-body proximity operations. The shared theme is coordinated multivariable learning under coupling; the mechanism and terminology are different [2602.14660].

A recurring misconception is that high-performance MIMO learning control necessarily requires either a full prior model or extensive manual tuning. The DILC formulation in [2509.18723] is explicitly intended to reject that premise: it is data-driven, self-parametrizing, and designed to operate without prior system knowledge or manual parameter tuning. A second misconception is that DILC is merely a rebranding of ordinary model-based ILC. The defining feature of DILC in [2509.18723] is the simultaneous update of the model estimate and the control input on every trial.

## 7. Significance and research outlook

Within the MIMO ILC literature, MIMO DILC represents a synthesis of autonomous model learning and autonomous control learning in the trial domain. Its reported advantages are plug-and-play deployment, compatibility with established ILC design functions, scalability to systems with many inputs and outputs, and validation on both simulated and real robotic systems [2509.18723].

Its significance is sharpened by the background problems identified elsewhere in the literature. Multivariable interaction can force trade-offs between user effort and performance [1806.08550]; nonsquare systems and nonrepetitive uncertainties can create contradictory convergence conditions for input and output analyses [1910.10305]; and purely model-free gradient estimation methods can require careful experimental design for efficiency and robustness [2111.08445]. MIMO DILC addresses a different but complementary axis of the problem by making model acquisition itself an iterative learning process coupled to tracking improvement [2509.18723].

A plausible implication is that future work will continue to connect these lines: autonomous DILC-type model/control co-learning, interaction-aware robust MIMO ILC design, and transformed-domain convergence analysis for nonsquare or uncertain systems. The available sources support a narrower, factual conclusion: MIMO DILC, as formulated in [2509.18723], is a data-driven, self-parametrized framework for simultaneous model learning and trial-domain tracking improvement in repetitive MIMO systems, with monotonic convergence conditions in the LTI case and experimental validation on high-dimensional robotic platforms.

Source: https://www.emergentmind.com/topics/mimo-dual-iterative-learning-control-dilc