{"id":"510461da-e7a9-49f7-bc79-0d6043920709","arxiv_id":"2607.06535","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":3,"one_line_summary":"A dual-pathway Neural-ESO uses a Lipschitz-bounded neural network for feedforward disturbance prediction and a conventional ESO for online correction, guaranteeing uniform ultimate boundedness of the closed-loop system.","lead":"This paper presents a control architecture that combines a neural network with a traditional Extended State Observer (ESO) to reject disturbances like ground effect during drone landing. A smart generalist might read it to understand how to safely integrate machine learning into safety-critical robotics without losing stability guarantees.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Theorem 1 omits the small-gain condition (Eq. 33) from its statement, yet the proof's UUB conclusion depends on it; the paper asserts satisfiability without proving it for the experimental system.","rationale":"The proof structure is sound: the error dynamics decomposition (Eqs. 12, 16), the interconnection bound (Lemma 1), and the composite Lyapunov analysis (Eqs. 24–34) follow standard and correct nonlinear control methodology. The Lipschitz bound on the neural network genuinely reduces the interconnection gains (γ_e = L_N λ_c, γ_ẽ = L_N), which is the paper's key insight and is correctly derived. The experimental results provide reasonable empirical support, and the L_N ablation (Table III) qualitatively confirms the predicted accuracy-robustness trade-off. However, the theorem statement is imprecise: it claims UUB follows from Assumptions 1–3 and the Lipschitz network, but the proof additionally requires the small-gain condition (33). This condition is not automatically satisfied — it depends on the relationship between L_N, the Lyapunov matrix norms (which are gain-dependent), and the free Young's inequality parameters. The paper asserts satisfiability without verification. This is common practice in control theory papers, but it leaves a gap between the stated theorem and what is actually proven. The concern is load-bearing because the paper's central contribution is the sufficiency claim: 'enforcing a Lipschitz bound guarantees UUB.' Without confirming that the small-gain condition is met (at least for the experimental system), this sufficiency claim is conditional rather than unconditional. A CONDITIONAL verdict reflects that the theoretical framework is correct but the central claim as stated is slightly stronger than what is rigorously proven. The reader's ACCEPT verdict is reasonable given that this is standard practice, but the gap is worth flagging. The reader's identified weakest assumption (Assumption 1, bounded disturbance derivative) is a legitimate domain limitation but is standard in ESO literature and not specific to this paper's contribution. The small-gain condition gap is more directly tied to the paper's central claim about sufficiency of the Lipschitz bound.","tokens_in":13884,"tokens_out":6893,"duration_ms":522188,"concrete_test":"Compute the small-gain condition (Eq. 33) numerically for the experimental quadrotor system using the actual deployed gains (k_p, k_d, ω_o) and L_N = 1.0. Specifically: solve the Lyapunov equations A_e^T P_e + P_e A_e = -Q_e and A_obs^T P_obs + P_obs A_obs = -Q_obs for the experimental gain values; compute c_o, c_e, ε_o, λ_e, λ_o; and verify whether λ_e > δ_e + c_o and λ_o > ε_o + c_e hold for some choice of δ_e, δ_o, δ_c > 0. If the condition cannot be satisfied with the deployed gains, the UUB guarantee does not apply to the experimental results, weakening the theory-experiment connection.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 1 states that UUB follows from Assumptions 1–3 and an L_N-Lipschitz network. However, the proof derives an explicit small-gain condition (Eq. 33): λ_e > δ_e + c_o and λ_o > ε_o + c_e, where c_o = (||P_obs B_obs|| · L_N λ_c)²/δ_o and ε_o includes the term 2||P_obs B_obs|| · L_N. These inequalities are not guaranteed by the Lipschitz bound alone — they require the control gains (k_p, k_d) and observer bandwidth (ω_o) to be sufficiently large relative to L_N and the Lyapunov matrix norms. The paper asserts (after Eq. 33) that 'the gains can be selected to satisfy the above bounds without requiring either high-gain control or high-gain observer,' but provides no constructive proof of existence, no numerical verification for the experimental quadrotor system, and no explicit relationship between L_N and the minimum required gains. This matters because the central claim — that enforcing a Lipschitz bound is 'sufficient' for UUB — is only true conditional on the small-gain condition being met. If the required gains for a given L_N are impractically large (e.g., exceeding actuator limits or amplifying noise), the guarantee is vacuous for real systems. The experimental results in Table III (L_N = 2.0 causes instability) are consistent with the small-gain condition being violated, but the paper does not confirm that L_N = 1.0 actually satisfies it for the deployed gains. The reader's concern about Assumption 1 (bounded disturbance derivative) is valid but standard in ESO literature; the gap between the theorem statement and its proof requirements is the more load-bearing issue for the paper's central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This letter proposes Neural-ESO, a dual-pathway disturbance-rejection architecture combining a Lipschitz-bounded neural network feedforward pathway with a conventional ESO corrective pathway. The authors derive a UUB stability guarantee via composite Lyapunov analysis and a small-gain condition, showing that the Lipschitz constant L_N governs the interconnection gain between tracking and observer error subsystems. The framework is validated on a quadrotor landing task under ground-effect disturbances, including in-distribution, OOD, and high-speed maneuver scenarios, with comparisons against PD, PD+ESO, PD+NN, and a neural-observer baseline. Code and data are publicly available.","tokens_in":14752,"tokens_out":1280,"duration_ms":275061,"significance":"The paper makes a solid contribution by providing an explicit, falsifiable connection between the neural network Lipschitz constant and closed-loop stability requirements via a small-gain condition (Eq. 33). The dual-pathway architecture is well-motivated, and the total disturbance retraining (TDR) mechanism for OOD adaptation is a practical and interesting design choice. The release of reproducible code and example datasets is a notable strength. The experimental validation across three scenarios with statistical reporting (5 trials each) and an ablation on L_N provides concrete evidence for the accuracy-robustness trade-off. The stability analysis correctly leverages standard tools (ISS, Young's inequality) and the interconnection gain quantification in Lemma 1 is clean.","major_comments":[{"comment":"§IV.D, Theorem 1 statement: The theorem states that UUB follows from Assumptions 1–3 and an L_N-Lipschitz network, but the proof derives an explicit small-gain condition (Eq. 33: λ_e > δ_e + c_o and λ_o > ε_o + c_e) that is not included in the theorem's hypotheses. Since c_o and ε_o both depend on L_N (through γ_e and γ_˜e), and the dissipation terms λ_e, λ_o depend on the control and observer gains, the small-gain condition is an additional load-bearing requirement beyond the Lipschitz bound alone. The theorem statement should be revised to explicitly include the small-gain condition (Eq. 33) as a hypothesis, or the proof should demonstrate that the condition is automatically satisfied for any choice of L_N and gains. As stated, the claim that enforcing a Lipschitz bound is 'sufficient' for UUB is only true conditional on the small-gain inequalities holding.","section":null},{"comment":"§IV.D, after Eq. (33): The paper asserts that 'the gains can be selected to satisfy the above bounds without requiring either high-gain control or high-gain observer,' but provides no constructive proof of existence, no numerical verification for the experimental quadrotor system, and no explicit relationship between L_N and the minimum required gains. This matters because the central sufficiency claim depends on the feasibility of satisfying Eq. (33). The experimental results in Table III (L_N = 2.0 causes instability) are consistent with the small-gain condition being violated, but the paper does not confirm that L_N = 1.0 actually satisfies it for the deployed gains (k_p, k_d, ω_o). Adding a numerical check—computing the terms in Eq. (33) for the experimental parameters—would substantially strengthen the claim and close the gap between theory and experiment.","section":null}],"minor_comments":[{"comment":"§IV.C, Eq. (22): The term λ_c is defined as obtained from control gains satisfying ||k_p e|| + ||k_d ė|| ≤ λ_c ||e_state||, but the explicit expression for λ_c in terms of k_p and k_d is not given. Providing this would aid reproducibility.","section":null},{"comment":"§IV.D, Eq. (32): The notation ε_o is introduced as 2c_pb γ_˜e + δ_o + δ_c, but the symbol ε_o does not appear in the final small-gain condition (33) in a way that makes the correspondence with λ_o > ε_o + c_e immediately clear. A brief clarifying sentence would help readers follow the derivation.","section":null},{"comment":"§V.F, Table III: The caption states results are across 5 trials, but the table does not specify which rows correspond to the nominal setup versus the OOD setup. Adding row labels or splitting the table would improve clarity.","section":null},{"comment":"§III.A.2: The network architecture is described as 4 fully connected layers with dimensions 5→20→25→10→1, but the input vector z_k in §V.C has 6 components (q_k, q̇_k, T_{k-1}, ϕ_k, θ_k). This discrepancy should be clarified—either the architecture description or the input dimension is incorrect.","section":null},{"comment":"§II.B, Eq. (6): The observer gains β_1, β_2, β_3 are stated to be chosen as diagonal matrices with eigenvalues at −ω_o, but the specific relationship between β_i and ω_o (e.g., β_1 = 3ω_o, β_2 = 3ω_o², β_3 = ω_o³ for a third-order ESO) is not given. Specifying this would aid reproducibility.","section":null},{"comment":"The video and code links in the abstract are a valuable addition. For the final version, please ensure the repository includes the specific L_N values and observer bandwidth ω_o used in experiments.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The paper appears to have been accepted to IEEE RA-L already (header states 'ACCEPTED JUNE, 2026'). If this review is for a post-acceptance revision or a separate venue, the major comments above should still be addressed as they concern the precision of the central theoretical claim. The stress-test concern about the small-gain condition omission from Theorem 1 is well-founded and is the most important issue to address. The reader's concern about Assumption 1 (bounded disturbance derivative) is valid but standard in the ESO/ADRC literature; I would not elevate it to a major comment."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for the careful reading and constructive feedback. Both major comments are well-taken and identify genuine gaps between the theorem statement and the proof, and between theory and experiment. We will revise the manuscript accordingly.","responses":[{"response":"The referee is correct. The small-gain inequalities in Eq. (33) — namely λ_e > δ_e + c_o and λ_o > ε_o + c_e — are load-bearing requirements in the proof: they ensure that the composite Lyapunov derivative is negative definite outside a bounded region. As the referee notes, c_o and ε_o both depend on L_N (through γ_e and γ_˜e), so the small-gain condition is not automatically satisfied by the Lipschitz bound alone; it also constrains the control and observer gains. The current theorem statement is therefore imprecise in claiming that the Lipschitz bound alone is 'sufficient' for UUB. We will revise Theorem 1 to explicitly include the small-gain condition (Eq. 33) as a hypothesis. The revised statement will read approximately as: 'Given Assumptions 1–3, an L_N-Lipschitz network f_N(z), and control/observer gains satisfying the small-gain condition (33), the closed-loop error state E is UUB.' We will also adjust the surrounding text (including the abstract and contributions) to state that the Lipschitz bound is necessary but not alone sufficient — it must be paired with gains satisfying the small-gain inequalities. This is a correction of the theorem statement, not a change to the proof itself, which already derives and uses Eq. (33).","revision_made":"yes","referee_comment":"§IV.D, Theorem 1 statement: The theorem states that UUB follows from Assumptions 1–3 and an L_N-Lipschitz network, but the proof derives an explicit small-gain condition (Eq. 33) that is not included in the theorem's hypotheses. The theorem statement should be revised to explicitly include the small-gain condition as a hypothesis."},{"response":"The referee raises a valid point. The assertion that the small-gain condition is feasible 'without requiring either high-gain control or high-gain observer' is currently unsupported by either a constructive argument or numerical verification. We agree that this gap should be addressed. In the revision, we will take the following steps: (1) We will add a numerical computation of the terms in Eq. (33) using the experimental quadrotor parameters (mass m_v, control gains k_p, k_d, observer bandwidth ω_o, and the trained network's L_N = 1.0) to verify that the small-gain inequalities are satisfied for the deployed configuration. This involves computing λ_e = λ_min(Q_e), λ_o = λ_min(Q_obs), and the interconnection gains c_o, c_e, ε_o, δ_e from the Lyapunov equation solutions and the expressions in the proof. (2) We will also compute these terms for L_N = 2.0 (which caused instability in Table III) to check whether the small-gain condition is violated, providing a concrete link between the theoretical condition and the experimental observation. (3) We will soften the claim about 'not requiring high-gain control or observer' to a more precise statement: the small-gain condition can be satisfied with moderate gains for sufficiently small L_N, and we will report the specific numerical margins for our system. We acknowledge that we cannot provide a fully constructive proof of gain feasibility for arbitrary systems and arbitrary L_N — the feasibility depends on system-specific quantities. The numerical verification for our experimental platform is the most honest and useful contribution we can make here.","revision_made":"yes","referee_comment":"§IV.D, after Eq. (33): The paper asserts that gains can be selected to satisfy the bounds without high-gain control or observer, but provides no constructive proof, no numerical verification for the experimental quadrotor, and no explicit relationship between L_N and minimum required gains. Adding a numerical check computing the terms in Eq. (33) for the experimental parameters would strengthen the claim."}],"tokens_in":13652,"tokens_out":1067,"duration_ms":86326,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The main thing to know: this paper combines a Lipschitz-bounded neural network feedforward path with a conventional ESO correction path, and proves UUB stability via a composite Lyapunov/small-gain argument. The architecture is sensible, the analysis is mostly clean, and the hardware experiments on a quadrotor are thorough. It deserves a serious referee who asks for one specific fix to the theorem statement.","headline":"Neural-ESO: dual-pathway learning-based ESO with Lyapunov UUB guarantee — solid contribution with a theorem statement that needs tightening","tokens_in":14718,"tokens_out":785,"would_cite":true,"duration_ms":42814,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Lipschitz-Bounded Neural Nets Keep Learning-Based Robot Control Stable","keywords":["learning-based control","extended state observer","Lipschitz constraint","uniform ultimate boundedness","small-gain theorem","disturbance rejection","dual-pathway architecture","spectral normalization"],"falsifier":"Deploy the Neural-ESO on a system where the disturbance derivative exceeds the assumed bound L_1 (e.g., a quadrotor encountering a sharp wind gust or ground-effect discontinuity), and verify whether the closed-loop error exceeds the predicted UUB bound or becomes unstable.","tokens_in":14050,"feed_emoji":"🤖","tokens_out":1175,"duration_ms":67457,"temperature":0.7,"pith_summary":"This paper proposes a dual-pathway control architecture called Neural-ESO that combines a neural network's feedforward disturbance prediction with a conventional Extended State Observer (ESO) for online correction. The central claim is that if the neural network is constrained to be Lipschitz-bounded—meaning its output cannot change faster than a fixed multiple of its input change—then the closed-loop error dynamics of the entire system are Uniformly Ultimately Bounded (UUB), meaning errors converge to and remain within a bounded region. The proof proceeds by decomposing the system into tracking-error and observer-error subsystems, showing each is Input-to-State Stable, quantifying their interconnection gains (which scale linearly with the network's Lipschitz constant), and verifying a small-gain condition. The architecture is validated on a quadrotor landing task under ground-effect disturbances, including out-of-distribution scenarios where the neural network's predictions become inaccurate but the ESO corrective pathway maintains stability.","feed_headline":"Lipschitz Neural Net + Classical Observer = Provably Stable Robot Control","feed_subtitle":"Constraining a neural network's sensitivity to a fixed bound lets a dual-pathway architecture guarantee bounded errors even when learning is","key_machinery":"The central mechanism is the dual-pathway disturbance decomposition: the total disturbance estimate is split as f_hat = f_N + Delta_f_hat, where f_N is the Lipschitz-bounded neural network prediction and Delta_f_hat is the ESO's online estimate of the residual. This decomposition ensures the tracking error is driven only by the residual estimation error, and the observer error dynamics are driven by the derivative of the residual, which is bounded via the Lipschitz property. The small-gain condition (inequalities 33) links L_N to required stabilization gains.","core_discovery":"The paper's core discovery is that enforcing a Lipschitz bound on the neural network learning component is a sufficient condition for closed-loop stability of a learning-based observer-controller system. Specifically, the interconnection gains between the tracking-error and observer-error subsystems are shown to be proportional to the Lipschitz constant L_N of the network, so by constraining L_N via spectral normalization and selecting appropriate controller/observer gains to satisfy a small-gain inequality, the composite Lyapunov function derivative becomes negative outside a bounded region, proving UUB. This transforms the Lipschitz constant into a single tuning knob that trades learning表达","pith_inferences":["The stability guarantee depends on the true disturbance derivative being bounded (Assumption 1), which may not hold for discontinuous aerodynamic phenomena like sudden wind gusts or sharp ground-effect transitions—this limits the class of disturbances for which the UUB guarantee is valid.","The small-gain inequalities (33) implicitly impose an upper bound on L_N relative to the controller and observer gains; if the network requires a large Lipschitz constant to approximate a highly nonlinear disturbance, the required stabilization gains may become impractically high.","The Total Disturbance Retraining mechanism assumes that the ESO compensation accurately reconstructs the true disturbance in the new domain; if the ESO bandwidth is insufficient for the new disturbance characteristics, the retraining labels will be noisy or biased.","The proof structure—decomposing into ISS subsystems and bounding interconnection gains via the Lipschitz property—could potentially extend to other function approximators (e.g., Gaussian processes, kernel methods) where a Lipschitz or gradient bound is available."],"forward_implications":["The Lipschitz constant L_N serves as a principled design parameter: decreasing it reduces required control/observer gains and improves OOD robustness, while increasing it improves in-distribution accuracy but demands higher stabilization gains.","The dual-pathway architecture enables safe data collection in new domains: even when the neural prior is inaccurate under OOD conditions, the ESO corrective pathway maintains stability, allowing the reconstructed total disturbance to be used as a training label for domain adaptation (Total Disturbance Retraining).","The stability guarantee holds during training transients when the neural network has not yet converged, addressing a gap in prior learning-based control methods that assume a converged model.","The framework applies to general Euler-Lagrange systems, suggesting transferability beyond quadrotors to other robotic platforms with lumped time-varying disturbances."],"fun_headline_variants":["Lipschitz-Bounded Neural Observer Guarantees Stable Disturbance Rejection","Constrain the Lipschitz Constant, Get Provably Bounded Tracking Error","Neural-ESO: One Tuning Knob Trades Learning Speed for Stability Proof","Dual-Pathway Control: Neural Feedforward + Classical ESO Correction = UUB","Spectral Normalization on Neural Controller Enables Lyapunov Stability Proof"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The proof requires that the time derivative of the true disturbance is bounded, which rules out discontinuous disturbances such as sudden wind gusts or abrupt aerodynamic transitions. If the real disturbance changes faster than this bound allows, the observer's input-to-state stability breaks down and the UUB guarantee no longer holds.","fun_headline_variants_meta":{"raw":{"variants":["Lipschitz-Bounded Neural Observer Guarantees Stable Disturbance Rejection","Constrain the Lipschitz Constant, Get Provably Bounded Tracking Error","Neural-ESO: One Tuning Knob Trades Learning Speed for Stability Proof","Dual-Pathway Control: Neural Feedforward + Classical ESO Correction = UUB","Spectral Normalization on Neural Controller Enables Lyapunov Stability Proof","Lipschitz Constant Becomes the Stability Knob for Learning-Based Observers","Neural Network + Classical Observer: Provably Bounded, Not Just Empirically Good","Quadrotor Landing Under Ground Effect: Lipschitz Bound Prevents Learning Failure"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":756,"prompt_tokens":486,"completion_tokens":270,"prompt_tokens_details":null},"tokens_in":486,"tokens_out":270,"duration_ms":10621,"temperature":1.0,"reasoning_tokens":139,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T02:23:57.183602+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"Deploy the Neural-ESO on a system where the disturbance derivative exceeds the assumed bound L_1 (e.g., a quadrotor encountering a sharp wind gust or ground-effect discontinuity), and verify whether the closed-loop error exceeds the predicted UUB bound or becomes unstable.","supporting_citations":[],"review_version":1}