pith. sign in

arxiv: 2604.21799 · v1 · submitted 2026-04-23 · 🧮 math.OC

H₂/H_(infty) Control for Stochastic Differential Systems with Partial Observation

Changwang Xiao , Nan Yang , Qingxin Meng This is my paper

Pith reviewed 2026-05-09 20:44 UTC · model grok-4.3

classification 🧮 math.OC
keywords H2/H∞ controlstochastic differential systemspartial observationNash equilibriumRiccati equationsKalman filterdifferential game
0
0 comments X

The pith

A Nash equilibrium for the H2/H∞ game exists precisely when the cross-coupled Riccati equations admit a solution, for linear stochastic systems under partial observations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper sets out to find conditions under which a controller can achieve both good average performance and strong robustness to disturbances in a linear stochastic system when the controller sees only noisy or incomplete measurements of the state. It frames the tradeoff as a two-player nonzero-sum game, applies the Kalman filter to estimate the state from observations, and derives a version of the bounded-real lemma that gives necessary and sufficient conditions for the robustness part. A reader might care because real-world control tasks, from aircraft to industrial processes, rarely provide perfect state information yet still need both efficiency and safety against worst-case inputs.

Core claim

The authors show that the H2/H∞ control problem for linear stochastic differential systems with partial observation can be reduced to the solvability of cross-coupled Riccati equations, whose solution yields the Nash equilibrium strategies for the associated differential game; this is obtained after deriving the Kalman filter equation and proving a stochastic bounded real lemma adapted to the partial-observation setting.

What carries the argument

Cross-coupled Riccati equations that characterize the Nash equilibrium of the nonzero-sum differential game after Kalman filtering of the partial observations.

Load-bearing premise

The system must be linear with additive white noise and the observations must be a linear function of the state plus independent noise.

What would settle it

For the UAV example, solve the cross-coupled Riccati equations and verify whether the resulting closed-loop trajectories meet both the target H2 cost and the H∞ bound under worst-case disturbances; a mismatch between solvability and achieved performance would refute the claimed equivalence.

Figures

Figures reproduced from arXiv: 2604.21799 by Changwang Xiao, Nan Yang, Qingxin Meng.

Figure 1
Figure 1. Figure 1: Trajectories of true states x(t) (blue solid lines) and estimated states xb(t) (red dashed lines) under the proposed controller and disturbance. The simulation results are presented in [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
read the original abstract

This paper investigates the $H_{2}/H_{\infty}$ control problem for linear stochastic differential systems under partial observation. Unlike existing studies that assume full state accessibility, we consider the scenario where the controller has access only to an observation process. The objective is to design a controller that balances the $H_2$ performance criterion with the $H_\infty$ robustness requirement under worst-case disturbances, formulated as a nonzero-sum differential game. Using the Kalman filtering method, we derive the corresponding optimal filtering equation. Furthermore, a Stochastic Bounded Real Lemma under the partial observation framework is established, providing necessary and sufficient conditions for the $H_\infty$ robustness constraint. We also show the connection between the existence of a Nash equilibrium and the solvability of the cross-coupled Riccati equations, and illustrate the effectiveness of the proposed approach through a numerical example involving an unmanned aerial vehicle (UAV).

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. This paper investigates the H_{2}/H_{∞} control problem for linear stochastic differential systems under partial observation. It formulates the problem as a nonzero-sum differential game, applies the Kalman filtering method to derive the optimal filtering equation, establishes a Stochastic Bounded Real Lemma under the partial observation framework to provide necessary and sufficient conditions for the H_{∞} robustness constraint, shows the connection between the existence of a Nash equilibrium and the solvability of cross-coupled Riccati equations, and illustrates the approach with a numerical UAV example.

Significance. If the central derivations hold, the work extends H_{2}/H_{∞} control results from full-state to partial-observation stochastic systems, which is relevant for applications such as UAV control where complete state access is unavailable. The explicit link between Nash equilibria and cross-coupled Riccati equations supplies a standard computational route, and the UAV example provides concrete validation of effectiveness.

major comments (2)
  1. [Kalman filtering derivation] The derivation applies the classical Kalman filter to obtain the estimator (see the filtering-equation section). However, the H_{∞} component is realized via an adversarial worst-case disturbance in the nonzero-sum game; this disturbance can correlate with the state and observation processes, violating the orthogonality principle that justifies the standard Kalman gain. No separate fixed-point argument or proof that the filter remains optimal under this correlation is supplied, which is load-bearing for the separation principle used throughout.
  2. [Stochastic Bounded Real Lemma] The Stochastic Bounded Real Lemma is asserted to give necessary and sufficient conditions for the H_{∞} constraint under partial observation. The lemma statement and its proof must explicitly verify that the innovation process remains suitable for the bounded-real property once the disturbance is allowed to depend on the estimation error; otherwise the necessity direction may fail.
minor comments (2)
  1. [Riccati equations and numerical example] Clarify the exact form of the cross-coupled Riccati equations (including any dependence on the filter gain) and confirm that the UAV example uses the same parameter values as the theoretical development.
  2. [Notation] Ensure all notation for the observation process y(t), filter state, and innovation is introduced consistently before first use.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. The major comments identify important technical points regarding the justification of the Kalman filter under adversarial disturbances and the necessity direction in the Stochastic Bounded Real Lemma. We address each point below and indicate the planned revisions.

read point-by-point responses

  1. Referee: The derivation applies the classical Kalman filter to obtain the estimator (see the filtering-equation section). However, the H_{∞} component is realized via an adversarial worst-case disturbance in the nonzero-sum game; this disturbance can correlate with the state and observation processes, violating the orthogonality principle that justifies the standard Kalman gain. No separate fixed-point argument or proof that the filter remains optimal under this correlation is supplied, which is load-bearing for the separation principle used throughout.

    Authors: We appreciate the referee highlighting this subtlety in applying the classical Kalman filter when the worst-case disturbance is determined endogenously by the game. In the manuscript the filter is derived from the linear system dynamics with the disturbance entering as an input whose effect is later optimized via the cross-coupled Riccati equations. The separation principle is invoked on the basis of the linear-quadratic structure. Nevertheless, an explicit argument confirming that the innovation process remains orthogonal to the estimation error under the equilibrium disturbance is not supplied. We will add a fixed-point verification or direct computation of the cross-covariance in the revised manuscript to close this gap. revision: yes

  2. Referee: The Stochastic Bounded Real Lemma is asserted to give necessary and sufficient conditions for the H_{∞} constraint under partial observation. The lemma statement and its proof must explicitly verify that the innovation process remains suitable for the bounded-real property once the disturbance is allowed to depend on the estimation error; otherwise the necessity direction may fail.

    Authors: The referee correctly notes that necessity in the partial-observation Stochastic Bounded Real Lemma requires the innovation to retain the martingale property even when the disturbance may depend on the estimation error. Our current proof establishes sufficiency via the Riccati solution and the quadratic form along the filtered trajectories. For necessity we have relied on the linear-Gaussian structure preserving whiteness of the innovation. We agree that an explicit check ruling out residual cross terms is needed. In the revision we will augment the lemma statement and its necessity proof with a direct verification that the innovation remains uncorrelated with the state estimate under the worst-case disturbance. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation grounded in standard Kalman filtering and Riccati solvability

full rationale

The paper applies the classical Kalman filter to obtain the optimal estimator under partial observations, establishes a Stochastic Bounded Real Lemma providing necessary and sufficient conditions for the H∞ constraint, and links Nash equilibrium existence to solvability of cross-coupled Riccati equations whose coefficients derive directly from the linear system matrices and filter gains. These steps follow standard stochastic control theory without reducing any central claim to a fitted parameter renamed as prediction, a self-citation chain, or an ansatz smuggled via prior work. The UAV numerical example is presented solely for illustration of effectiveness, with no indication that Riccati solutions are fitted to example data or that predictions are forced by construction. The derivation chain remains self-contained against external benchmarks in linear stochastic systems.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The approach rests on standard linear stochastic dynamics and Kalman-filter optimality, which are imported from prior literature rather than re-derived; the new elements are the partial-observation lemma and the game formulation.

axioms (2)
  • domain assumption The underlying system is a linear stochastic differential equation driven by white noise with a linear observation process.
    Invoked to justify direct use of the Kalman filter and extension of the bounded-real lemma.
  • ad hoc to paper A Nash equilibrium exists if and only if the cross-coupled Riccati equations admit a solution.
    This equivalence is asserted as the bridge between the game and the controller design.

pith-pipeline@v0.9.0 · 5453 in / 1339 out tokens · 28602 ms · 2026-05-09T20:44:47.321915+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages

  1. [1]

    LQG control with an H∞ performance bound: A Riccati equation approach,

    D. S. Bernstein and W. M. Haddad, “LQG control with an H∞ performance bound: A Riccati equation approach,” IEEE Transactions on Automatic Control, vol. 34, no. 3, pp. 293–305, 1989

    work page 1989

  2. [2]

    State-space solutions to standard H2 and H∞ control prob- lems,

    J. C. Doyle, K. Glover, P. P. Khargonekar, and B. A. Francis, “State-space solutions to standard H2 and H∞ control prob- lems,” IEEE Transactions on Automatic Control, vol. 34, no. 8, pp. 831–847, 1989

    work page 1989

  3. [3]

    A Nash game approach to mixed H2/H∞ control,

    D. J. N. Limebeer, B. D. O. Anderson, and B. Hendel, “A Nash game approach to mixed H2/H∞ control,” IEEE Transactions on Automatic Control, vol. 39, no. 1, pp. 69–82, 1994

    work page 1994

  4. [4]

    Stochastic H2/H∞ control with state-dependent noise,

    B. S. Chen and W. Zhang, “Stochastic H2/H∞ control with state-dependent noise,” IEEE Transactions on Automatic Con- trol, vol. 49, no. 1, pp. 45–57, 2004

    work page 2004

  5. [5]

    Stochastic H2/H∞ control for discrete-time systems with state and disturbance dependent noise,

    W. Zhang, Y. Huang, and H. Zhang, “Stochastic H2/H∞ control for discrete-time systems with state and disturbance dependent noise,” Automatica, vol. 43, no. 3, pp. 513–521, 2007

    work page 2007

  6. [6]

    Infinite horizon H2/H∞ optimal control for discrete-time Markov jump systems with (x, u, v)- dependent noise,

    T. Hou, W. Zhang, and H. Ma, “Infinite horizon H2/H∞ optimal control for discrete-time Markov jump systems with (x, u, v)- dependent noise,” Journal of Global Optimization, vol. 57, pp. 1245–1262, 2013

    work page 2013

  7. [7]

    Mixed H2/H∞ control of delayed Markov jump linear systems,

    W. Mei, C. Zhao, M. Ogura, and K. Sugimoto, “Mixed H2/H∞ control of delayed Markov jump linear systems,” IET Control Theory and Applications, vol. 14, no. 15, pp. 2076–2083, 2020

    work page 2076

  8. [8]

    Stochastic H2/H∞ control of nonlinear systems with time-delay and state-dependent noise,

    M. Gao, L. Sheng, and W. Zhang, “Stochastic H2/H∞ control of nonlinear systems with time-delay and state-dependent noise,” Applied Mathematics and Computation, vol. 266, pp. 429–440, 2015

    work page 2015

  9. [9]

    Mixed H2/H∞ control for discrete-time systems with input delay,

    X. Li, J. Xu, W. Wang, and H. Zhang, “Mixed H2/H∞ control for discrete-time systems with input delay,” IET Control Theory and Applications, vol. 12, no. 16, pp. 2221–2231, 2018

    work page 2018

  10. [10]

    Stochastic H2/H∞ Control for Mean-Field Stochastic Differential Systems with (x, u, v)-Dependent Noise,

    M. Wang, Q. Meng, Y. Shen, and P. Shi, “Stochastic H2/H∞ Control for Mean-Field Stochastic Differential Systems with (x, u, v)-Dependent Noise,” Journal of Optimization Theory and Applications, vol. 197, pp. 1024–1060, 2023

    work page 2023

  11. [11]

    H2/H∞ Control for Continuous-Time Mean-Field Stochastic Systems with Affine Terms,

    X. Fang, J. Moon, M. Tang, and Q. Meng, “ H2/H∞ Control for Continuous-Time Mean-Field Stochastic Systems with Affine Terms,” Control Theory & Applications, vol. 43, pp. 1-10, 2026

    work page 2026

  12. [12]

    Mixed H2/H∞ control for autonomous quadrotor landing via linear matrix inequality optimization,

    J. Qi and L. C. Zhao, “Mixed H2/H∞ control for autonomous quadrotor landing via linear matrix inequality optimization,” Journal of Aircraft, vol. 62, no. 5, pp. 1-12, 2025. JOURNAL OF XXXXXXXXXXXXX„ VOL. XX, NO. X, AUGUST 20XX 13

    work page 2025

  13. [13]

    Robust H∞ dual cascade MPC-based attitude control study of a quadcopter UA V,

    N. Hui, Y. Guo, X. Han, and B. Wu, “Robust H∞ dual cascade MPC-based attitude control study of a quadcopter UA V,” Actuators, vol. 13, no. 10, pp. 392, 2024

    work page 2024

  14. [14]

    Nonlinear control of quadrotor trajectory with discrete H∞,

    M. Hasanlu and M. Siavashi, “Nonlinear control of quadrotor trajectory with discrete H∞,” Journal of Mechanical Engineer- ing, Automation and Control Systems, vol. 6, no. 1, pp. 1-13, 2025

    work page 2025

  15. [15]

    On the separation theorem of stochastic control,

    W. M. Wonham, “On the separation theorem of stochastic control,” SIAM Journal on Control, vol. 6, no. 2, pp. 312–326, 1968

    work page 1968

  16. [16]

    M. H. A. Davis, Linear Estimation and Stochastic Control, Chapman and Hall, London, 1977

    work page 1977

  17. [17]

    Bensoussan, Stochastic Control of Partially Observable Sys- tems, Cambridge University Press, Cambridge, 1992

    work page 1992

  18. [18]

    Baras, and Robert J

    James, Matthew R., John S. Baras, and Robert J. Elliott. ”Risk-sensitive control and dynamic games for partially observed discrete-time nonlinear systems,” IEEE transactions on auto- matic control, vol. 39, no. 4, pp. 780-792, 2002

    work page 2002

  19. [19]

    Huang, G

    J. Huang, G. Wang, and J. Xiong, ”A maximum principle for partial information backward stochastic control problems with applications,” SIAM journal on Control and Optimization, vol. 48, no. 4, pp. 2106-2117, 2009

    work page 2009

  20. [20]

    Øksendal, Bernt, and Agnes Sulem, ”Maximum principles for optimal control of forward-backward stochastic differential equa- tions with jumps,” SIAM Journal on Control and Optimization, vol. 48, no. 5, pp. 2945-2976, 2010

    work page 2010

  21. [21]

    A linear-quadratic optimal control problem of forward-backward stochastic differential equa- tions with partial information,

    G. Wang, Z. Wu, and J. Xiong, “A linear-quadratic optimal control problem of forward-backward stochastic differential equa- tions with partial information,” IEEE Transactions on Automatic Control, vol. 60, no. 11, pp. 2904–2916, 2015

    work page 2015

  22. [22]

    Stochastic Linear-Quadratic Optimal Control with Partial Observation,

    J. Sun and J. Xiong, “Stochastic Linear-Quadratic Optimal Control with Partial Observation,” SIAM Journal on Control and Optimization, vol. 61, no. 3, pp. 1231–1247, 2023

    work page 2023

  23. [23]

    Separation Principle for Partially Ob- served Linear-Quadratic Optimal Control for Mean-Field Type Stochastic Systems,

    J. Moon and T. Basar, “Separation Principle for Partially Ob- served Linear-Quadratic Optimal Control for Mean-Field Type Stochastic Systems,” IEEE Transactions on Automatic Control, vol. 69, no. 12, pp. 8370–8385, 2024

    work page 2024

  24. [24]

    Maximum principles for forward-backward stochastic control systems with correlated state and observation noises,

    G. Wang, Z. Wu, and J. Xiong, “Maximum principles for forward-backward stochastic control systems with correlated state and observation noises,” SIAM Journal on Control and Optimization, vol. 51, no. 1, pp. 491–524, 2013

    work page 2013

  25. [25]

    Linear-quadratic partially observed forward-backward stochastic differential games and its applica- tion in finance,

    Z. Wu and Y. Zhuang, “Linear-quadratic partially observed forward-backward stochastic differential games and its applica- tion in finance,” Applied Mathematics and Computation, vol. 321, pp. 577–592, 2018

    work page 2018

  26. [26]

    The maximum principle for partially observed optimal control problems of mean-field FBSDEs,

    R. Li and F. Fu, “The maximum principle for partially observed optimal control problems of mean-field FBSDEs,” International Journal of Control, vol. 92, no. 10, pp. 2463–2472, 2019

    work page 2019

  27. [27]

    A partial information linear-quadratic optimal control problem of backward stochastic differential equation with its applications,

    P. Huang, G. Wang, and H. Zhang, “A partial information linear-quadratic optimal control problem of backward stochastic differential equation with its applications,” Science China Infor- mation Sciences, vol. 63, pp. 1–13, 2020

    work page 2020

  28. [28]

    Linear quadratic control of backward stochastic differential equation with partial informa- tion,

    G. Wang, W. Wang, and Z. Yan, “Linear quadratic control of backward stochastic differential equation with partial informa- tion,” Applied Mathematics and Computation, vol. 403, 126164, 2021

    work page 2021

  29. [29]

    Mixed H2/H∞ Con- trol With Event-Triggered Mechanism for Nonlinear Stochastic Systems With Closed-Loop Stackelberg Games,

    Z. Ming, H. Zhang, J. Zhang, and Y. Luo, “Mixed H2/H∞ Con- trol With Event-Triggered Mechanism for Nonlinear Stochastic Systems With Closed-Loop Stackelberg Games,” IEEE Transac- tions on Systems, Man, and Cybernetics: Systems, vol. 53, no. 10, pp. 6365–6374, 2023

    work page 2023

  30. [30]

    Hybrid H2/H∞ control of quadrotor based on model parameter uncertainty,

    C. Shi, D. Lin, B. Ren, and L. Yu, “Hybrid H2/H∞ control of quadrotor based on model parameter uncertainty,” Journal of Sichuan University of Science and Engineering(Natural Science Edition), vol. 31, pp. 25–33, 2018

    work page 2018

  31. [31]

    Kallianpur, Stochastic Filtering Theory, Springer, Berlin, 1980

    G. Kallianpur, Stochastic Filtering Theory, Springer, Berlin, 1980

    work page 1980

  32. [32]

    Yong and X

    J. Yong and X. Y. Zhou, Stochastic Controls: Hamiltonian Systems and HJB Equations, Springer, Berlin, 1999

    work page 1999