arXiv subjects

Xinfan Lin

Publications and source records attributed to Xinfan Lin.

At least 19 recordsLinked to original sources

arXiv2026

Exact Gaussian Moment Matching for Residual Networks: a Second-Order Method

We study the problem of propagating the mean and covariance of a general multivariate Gaussian distribution through a deep (residual) neural network using layer-by-layer moment matching. We close a longstanding gap by deriving exact moment matching for the probit, GeLU, ReLU (as a limit of GeLU), Heaviside (as a limit of probit), and sine activation functions; for both feedforward and generalized residual layers. On large-variance random networks, median KL-divergence improvements over common baselines range from tens to thousands of times. On a variational Bayes neural network, we show that our method improves upon a state-of-the-art deterministic inference method by an order of magnitude in KL divergence from Monte Carlo ground truth. We also give a smooth-distance error bound showing that, under regularity assumptions, moment matching removes the leading low-variance errors and propagates higher-order local accuracy through the layers of a network.

cs.LG↗

arXiv2026

Mass Scaling Laws for Energy-Optimal Horizontal Multirotor Flight

Multicopters are a type of aerial robot seeing extensive real-world applications but with range and endurance as critical bottlenecks. Optimization of its energy efficiency through energy-aware flight planning (e.g. cruise speed-scheduling) has hence been an important topic. The mass of the multicopter is an important factor to consider, as it has a significant impact on efficiency, and many common applications are subject to frequent mass variation due to payload change. This work derives closed-form scaling relationships between the energy-optimal steady horizontal flight behavior of a multicopter and its vehicle mass, and validates them experimentally on two multirotor platforms. Specifically, for mass-only variation, i.e., when payload changes the vehicle mass while the airframe geometry, aerodynamic characteristics, and rotor configuration remain unchanged, we show through a combination of numerical simulation, theoretical derivation, and experimental validation that (1) the energy-optimal velocity is proportional to the square root of vehicle mass; (2) the energy-optimal pitch angle is constant with respect to mass; and (3) the optimal energy per meter is proportional to mass. The analysis shows that these three relationships arise from a common mass-similarity property of the steady-state flight and rotor-aerodynamic equations. Furthermore, the obtained scaling relationships enable simple yet effective cruise-speed planning to optimize energy efficiency and extend the range and endurance of the multicopter.

cs.RO↗

arXiv2026

System Identification for Nonlinear Stochastic State-Space Models Using Ensemble Kalman Inversion with Application to Lithium-Ion Batteries

Nonlinear state-space system identification has attracted decades of interest, yet remains challenging because of the interplay among unknown parameters, nonlinear dynamics, and latent state variables. Its difficulty increases dramatically when the state and parameter dimensions increase. To address this challenge, we develop an ensemble Kalman inversion (EnKI)-based approach. This approach leverages EnKI to perform maximum a posteriori parameter estimation through successive local Gaussian approximations, enabling an iterative and incremental search for unknown parameters. The search combines Monte Carlo sampling with Kalman-type updates to evolve an ensemble of parameter samples, thereby offering empirical stability and the ability to handle strongly nonlinear models. The EnKI search is coupled with state estimation conditioned on the latest parameter estimates. This enables parameter identification along with latent state recovery. To improve scalability for large datasets, we further develop a cascaded EnKI (CasEnKI) method, which assimilates subdatasets sequentially to accelerate computation. We validate the proposed methods using a lithium-ion battery model with ten state variables and 18 unknown parameters. Simulation results show that the EnKI approach offers high identification accuracy and empirical stability in parameter search. Further, experimental results illustrate the computational efficiency and scalability of the CasEnKI method for large battery datasets.

eess.SY↗

arXiv2026

Stable initialization without the CLT

Successful training of deep neural networks is highly dependent on the distribution of the initial weights. If the weights are too large, network training blows up; if they are too small, the model fails to learn features. Stable initialization is the optimal moderation between these two extremes. The conventional theory of random networks uses the Central Limit Theorem to control inter-neuron dependencies, which introduces distributional approximation error and coupling between layers. For networks with sine activations, we derive the uniform-phase initialization, which obviates distributional approximation and fully decouples the layers. Ours is the first work to use the sine function's periodic symmetry. Models trained with the uniform-phase initialization outperform the state of the art in neural representation tasks like image and audio fitting. We find that our untuned models are competitive with the best-tuned baselines from previous work and support $μ$P width scaling.

cs.LG↗

arXiv2026

Equivalent Circuit Model based Electric Vehicle Evacuation with Mobile Charging Stations

The increasing penetration of electric vehicles (EVs) introduces new challenges for emergency evacuation planning due to limited driving range, long charging times, and constrained charging infrastructure, particularly under disaster induced disruptions. This paper proposes a novel optimization based evacuation framework for EVs using Equivalent Circuit Models (ECMs) to jointly address routing, charging, and congestion management. By leveraging electrical analogies, traffic flow is modeled as electrical current, travel time as resistance, and driving range as voltage, enabling the use of Kirchhoff laws to enforce flow balance and energy feasibility constraints. The proposed controllable ECM incorporates binary switches to regulate route selection and explicitly models charging delays and range replenishment at both Fixed Charging Stations (FCSs) and Mobile Charging Stations (MCSs). The resulting formulation leads to an integer programming problem that determines optimal evacuation routes, charging durations, and the placement and number of MCSs to minimize evacuation time. The framework is extended to multiple origin destination pairs using the principle of superposition and supports fairness aware performance metrics, including worst case, average, and variance based evacuation times. Simulation studies on large scale transportation networks in California demonstrate that the proposed approach significantly improves evacuation efficiency and robustness, particularly in scenarios with limited charging access, highlighting the critical role of MCSs in EV based emergency evacuations.

eess.SY↗

arXiv2026

Assumed Density Filtering and Smoothing with Neural Network Surrogate Models

The Kalman filter and Rauch-Tung-Striebel (RTS) smoother are optimal for state estimation in linear dynamic systems. With nonlinear systems, the challenge consists in how to propagate uncertainty through the state transitions and output function. For the case of a neural network model, we enable accurate uncertainty propagation using a recent state-of-the-art analytic formula for computing the mean and covariance of a deep neural network with Gaussian input. We argue that cross entropy is a more appropriate performance metric than RMSE for evaluating the accuracy of filters and smoothers. We demonstrate the superiority of our method for state estimation on a stochastic Lorenz system and a Wiener system, and find that our method enables more optimal linear quadratic regulation when the state estimate is used for feedback. Code available at https: //github.com/simontheflutist/analytic-moments.

eess.SY↗

arXiv2026

Instrumental variables system identification with $L^p$ consistency

Instrumental variables (eliminate the bias that afflicts least-squares identification of dynamical systems through noisy data, yet traditionally relies on external instruments that are seldom available for nonlinear time series data. We propose an IV estimator that synthesizes instruments from the data. We establish finite-sample $L^{p}$ consistency for all $p \ge 1$ in both discrete- and continuous-time models, recovering a nonparametric $\sqrt{n}$-convergence rate. On a forced Lorenz system our estimator reduces parameter bias by 200x (continuous-time) and 500x (discrete-time) relative to least squares and reduces RMSE by up to tenfold. Because the method only assumes that the model is linear in the unknown parameters, it is broadly applicable to modern sparsity-promoting dynamics learning models.

stat.ME↗

arXiv2026

Demystifying Lipschitz verification: positive matrices, negative results

The global Lipschitz constant of a neural network is related to robustness and generalization, yet unlike in many classical models, it is not plainly legible from the parameters. This has motivated sophisticated verification algorithms, especially semidefinite programming (SDP) based on incremental quadratic constraints on the activation functions, to improve on the fast but often loose product of layerwise Lipschitz constants (the trivial bound). We ask why Lipschitz verification is a problem in the first place. Our answer is that the difficulty is structural: estimating a network's Lipschitz constant requires knowing which hidden states are reachable, and reachability is NP-hard. If P!=NP, then reachability is a barrier to any polynomial-time algorithm. Through explicit constructions, we show that this blindness can force SDP-based bounds to inherit the same qualitative failures as the trivial bound, including but not limited to polynomial per-layer conservatism. We show that the difficulties of NP-hard questions are not isolated to worst-case computational reductions, but actually afflict every instance of the verification problem. Thus SDP is not sufficient for Lipschitz verification. We also argue that it is not necessary: several apparent failures of the trivial bound arise from removable parameterization pathologies, and can be mitigated by optimizing or regularizing the trivial bound itself. We demonstrate this claim via a "spherical cow" linear model and numerical proofs of concept. While the main contribution is theoretical and negative, we finally motivate a novel form of trigonometric layers that do not need biases for universal approximation. Combined with trivial bound regularization, they make the trivial bound provably and practically tight.

cs.LG↗

arXiv2026

Incremental stability in $p=1$ and $p=\infty$: classification and synthesis

All Lipschitz dynamics with the weak infinitesimal contraction (WIC) property can be expressed as a Lipschitz nonlinear system in proportional negative feedback -- this statement, a ``structure theorem,'' is true in the $p=1$ and $p=\infty$ norms. Equivalently, a Lipschitz vector field is WIC if and only if it can be written as a scalar decay plus a Lipschitz-bounded residual. We put this theorem to use using neural networks to approximate Lipschitz functions. This results in a map from unconstrained parameters to the set of WIC vector fields, enabling standard gradient-based training with no projections or penalty terms. Because the induced $1$- and $\infty$-norms of a matrix reduce to row or column sums, Lipschitz certification costs only $O(d^2)$ operations -- the same order as a forward pass and appreciably cheaper than eigenvalue or semidefinite methods for the $2$-norm. Numerical experiments on a planar flow-fitting task and a four-node opinion network demonstrate that the parameterization (re-)constructs contracting dynamics from trajectory data. In a discussion of the expressiveness of non-Euclidean contraction, we prove that the set of $2\times 2$ systems that contract in a weighted $1$- or $\infty$-norm is characterized by an eigenvalue cone, a strict subset of the Hurwitz region that quantifies the cost of moving away from the Euclidean norm.

eess.SY↗

arXiv2026

Accelerating Bayesian Optimization for Nonlinear State-Space System Identification with Application to Lithium-Ion Batteries

This paper studies system identification for nonlinear state-space models, a problem that arises across many fields yet remains challenging in practice. Focusing on maximum likelihood estimation, we employ Bayesian optimization (BayesOpt) to address this problem by leveraging its derivative-free global search capability enabled by surrogate modeling of the likelihood function. Despite these advantages, standard BayesOpt often suffers from slow convergence, high computational cost, and practical difficulty in attaining global optima under limited computational budgets, especially for high-dimensional nonlinear models with many unknown parameters. To overcome these limitations, we propose an accelerated BayesOpt framework that integrates BayesOpt with the Nelder--Mead method. Heuristics-based, the Nelder--Mead method provides fast local search, thereby assisting BayesOpt when the surrogate model lacks fidelity or when over-exploration occurs in broad parameter spaces. The proposed framework incorporates a principled strategy to coordinate the two methods, effectively combining their complementary strengths. The resulting hybrid approach significantly improves both convergence speed and computational efficiency while maintaining strong global search performance. In addition, we leverage an implicit particle filtering method to enable accurate and efficient likelihood evaluation. We validate the proposed framework on the identification of the BattX model for lithium-ion batteries, which features ten state dimensions, 18 unknown parameters, and strong nonlinearity. Both simulation and experimental results demonstrate the effectiveness of the proposed approach as well as its advantages over alternative methods.

eess.SY↗

arXiv2026

Machine Learning Detection of Lithium Plating in Lithium-ion Cells: A Gaussian Process Approach

Lithium plating during fast charging is a critical degradation mechanism that accelerates capacity fade and can trigger catastrophic safety failures. Recent work has shown that plating onset can manifest in incremental-capacity analysis as an additional high-voltage feature above 4.0 V, often appearing as a secondary peak or shoulder distinct from the main intercalation peak complex; however, conventional methods for computing dQ/dV rely on finite differencing with filtering, which amplifies sensor noise and introduces bias in feature location. In this paper, we propose a Gaussian Process (GP) framework for lithium plating detection by directly modeling the charge-voltage relationship Q(V) as a stochastic process with calibrated uncertainty. Leveraging the property that derivatives of GPs remain GPs, we infer dQ/dV analytically and probabilistically from the posterior, enabling robust detection without ad hoc smoothing. The framework provides three key benefits: (i) noise-aware inference with hyperparameters learned from data, (ii) closed-form derivatives with credible intervals for uncertainty quantification, and (iii) scalability to online variants suitable for embedded BMS. Experimental validation on Li-ion coin cells across a range of C-rates (0.2C-1C) and temperatures (0-40$^\circ$C) demonstrates that the GP-based method reliably resolves distinct high-voltage secondary peak features under low-temperature, high-rate charging, while correctly reporting no features in non-plating cases. The concurrence of GP-identified differential features, reduced charge throughput, capacity fade measured via reference performance tests, and post-mortem microscopy confirmation supports the interpretation of these signatures as plating-related, establishing a practical pathway for real-time lithium plating detection.

cs.LG↗

arXiv2025

Joint Parameterization of Hybrid Physics-Based and Machine Learning Li-Ion Battery Model

Electrochemical hybrid battery models have major potential to enable advanced physics-based control, diagnostic, and prognostic features for next-generation lithium-ion battery management systems. This is due to the physical significance of the electrochemical model, which is complemented by a machine learning model that compensates for output prediction errors caused by system uncertainties. While hybrid models have demonstrated robust output prediction performance under large system uncertainties, they are highly susceptible to the influence of uncertainties during parameter identification, which can compromise the physical significance of the model. To address this challenge, we present a parameter estimation framework that explicitly considers system uncertainties through a discrepancy function. The approach also incorporates a downsampling procedure to address the computational barriers associated with large time-series data sets, as are typical in the battery domain. The framework was validated in simulation, yielding several mean parameter estimation errors that were one order of magnitude smaller than those of the conventional least squares approach. While developed for the high-uncertainty, electrochemical hybrid modeling context, the estimation framework is applicable to all models and is presented in a generalized form.

eess.SY↗

arXiv2025

Robust Estimation of Battery State of Health Using Reference Voltage Trajectory

Accurate estimation of state of health (SOH) is critical for battery applications. Current model-based SOH estimation methods typically rely on low C-rate constant current tests to extract health parameters like solid phase volume fraction and lithium-ion stoichiometry, which are often impractical in real-world scenarios due to time and operational constraints. Additionally, these methods are susceptible to modeling uncertainties that can significantly degrade the estimation accuracy, especially when jointly estimating multiple parameters. In this paper, we present a novel reference voltage-based method for robust battery SOH estimation. This method utilizes the voltage response of a battery under a predefined current excitation at the beginning of life (BOL) as a reference to compensate for modeling uncertainty. As the battery degrades, the same excitation is applied to generate the voltage response, which is compared with the BOL trajectory to estimate the key health parameters accurately. The current excitation is optimally designed using the Particle Swarm Optimization algorithm to maximize the information content of the target parameters. Simulation results demonstrate that our proposed method significantly improves parameter estimation accuracy under different degradation levels, compared to conventional methods relying only on direct voltage measurements. Furthermore, our method jointly estimates four key SOH parameters in only 10 minutes, making it practical for real-world battery health diagnostics, e.g., fast testing to enable battery repurposing.

eess.SY↗

arXiv2025

Debiasing Continuous-time Nonlinear Autoregressions

We study how to identify a class of continuous-time nonlinear systems defined by an ordinary differential equation affine in the unknown parameter. We define a notion of asymptotic consistency as $(n, h) \to (\infty, 0)$, and we achieve it using a family of direct methods where the first step is differentiating a noisy time series and the second step is a plug-in linear estimator. The first step, differentiation, is a signal processing adaptation of the nonparametric statistical technique of local polynomial regression. The second step, generalized linear regression, can be consistent using a least squares estimator, but we demonstrate two novel bias corrections that improve the accuracy for finite $h$. These methods significantly broaden the class of continuous-time systems that can be consistently estimated by direct methods.

eess.SY↗

arXiv2025

A New Type of Nonlinear Disturbance Rejection

Asymptotic disturbance rejection (equivalently tracking) for nonlinear systems has been studied only in qualitative terms (the state is asymptotically stable under bounded disturbances). We show how to prove quantitative performance guarantees for the nonlinear servomechanism problem. Our technique originates by applying a gain inequalities point of view to an ad fontes reexamination of the linear problem: what is the nonlinear equivalent of a sensitivity transfer function with a zero at the origin? We answer: a nonlinear input-output system is high-pass if its output is stable with respect to the \emph{derivative} of the input. We first show that definition generalizes high-pass resistor-capacitor circuit analysis to accommodate nonlinear resistors. We then show that this definition generalizes the steady-state disturbance rejection property of integral feedback controllers for linear systems. The theoretical payoff is that low-frequency disturbance rejection is captured by a quantitative, non-asymptotic output cost bound. Finally, we raise theoretical questions about compositionality of nonlinear operators.

eess.SY↗

arXiv2025

Enhancing Multirotor Drone Efficiency: Exploring Minimum Energy Consumption Rate of Forward Flight under Varying Payload

Multirotor unmanned aerial vehicle is a prevailing type of aircraft with wide real-world applications. Energy efficiency is a critical aspect of its performance, determining the range and duration of the missions that can be performed. In this study, we show both analytically and numerically that the optimum of a key energy efficiency index in forward flight, namely energy per meter traveled per unit mass, is a constant under different vehicle mass (including payload). Note that this relationship is only true under the optimal forward velocity that minimizes the energy consumption (under different mass), but not under arbitrary velocity. The study is based on a previously developed model capturing the first-principle energy dynamics of the multirotor, and a key step is to prove that the pitch angle under optimal velocity is a constant. By employing both analytical derivation and validation studies, the research provides critical insights into the optimization of multirotor energy efficiency, and facilitate the development of flight control strategies to extend mission duration and range.

cs.RO↗

arXiv2024

Energy-Optimal Planning of Waypoint-Based UAV Missions -- Does Minimum Distance Mean Minimum Energy?

Multirotor unmanned aerial vehicle is a prevailing type of aerial robots with wide real-world applications. The energy efficiency of the robot is a critical aspect of its performance, determining the range and duration of the missions that can be performed. This paper studies the energy-optimal planning of the multirotor, which aims at finding the optimal ordering of waypoints with the minimum energy consumption for missions in 3D space. The study is performed based on a previously developed model capturing first-principle energy dynamics of the multirotor. We found that in majority of the cases (up to 95%) the solutions of the energy-optimal planning are different from those of the traditional traveling salesman problem which minimizes the total distance. The difference can be as high as 14.9%, with the average at 1.6%-3.3% and 90th percentile at 3.7%-6.5% depending on the range and number of waypoints in the mission. We then identified and explained the key features of the minimum-energy order by correlating to the underlying flight energy dynamics. It is shown that instead of minimizing the distance, coordination of vertical and horizontal motion to promote aerodynamic efficiency is the key to optimizing energy consumption.

cs.RO↗

arXiv2024

Structural Constraints for Physics-augmented Learning

When the physics is wrong, physics-informed machine learning becomes physics-misinformed machine learning. A powerful black-box model should not be able to conceal misconceived physics. We propose two criteria that can be used to assert integrity that a hybrid (physics plus black-box) model: 0) the black-box model should be unable to replicate the physical model, and 1) any best-fit hybrid model has the same physical parameter as a best-fit standalone physics model. We demonstrate them for a sample nonlinear mechanical system approximated by its small-signal linearization.

cs.LG↗