An official website of the United States government
Abstract Navigating complex terrains requires dynamic interactions between the substrate, musculoskeletal, and sensorimotor systems. Current perturbation studies have mostly used visible terrain height perturbations, which do not allow us to distinguish among the neuromechanical contributions of feedforward control, feedback-mediated, and mechanical perturbation responses. Here, we use treadmill-belt speed perturbations to induce a targeted perturbation to foot speed only, and without terrain-induced changes in joint posture and leg loading at stance onset. Based on previous studies suggesting a proximo-distal gradient in neuromechanical control, we hypothesized that distal joints would exhibit larger changes in joint kinematics, compared to proximal joints. Additionally, we expected birds to use feedforward strategies to increase the intrinsic stability of gait. To test these hypotheses, seven adult guinea fowl were video recorded while walking on a motorized treadmill, during both steady and perturbed trials. Perturbations consisted of repeated exposures to a deceleration and acceleration of the treadmill-belt speed. Surprisingly, we found that joint angular trajectories and center of mass fluctuations remain very similar, despite substantial perturbation of foot velocity by the treadmill belt. Hip joint angular trajectories exhibit the largest changes, with the birds adopting a slightly more flexed position across all perturbed strides. Additionally, we observed increased stride duration across all strides, consistent with feedforward changes in the control strategy. The speed perturbations mainly influenced the timing of stance and swing, with the largest kinematic changes in the strides directly following a deceleration. Our findings do not support the general hypothesis of a proximo-distal gradient in joint control, as distal joint kinematics remain largely unchanged. Instead, we find that leg angular trajectory and the timing of stance and swing are most sensitive to this specific perturbation, and leg length actuation remains largely unchanged. Our results are consistent with modular task-level control of leg length and leg angle actuation, with different neuromechanical control and perturbation sensitivity in each actuation mode. Distal joints appear to be sensitive to changes in vertical loading but not foot fore-aft velocity. Future directions should include in vivo studies of muscle activation and force–length dynamics to provide more direct evidence of the sensorimotor control strategies for stability in response to belt-speed perturbations.
more »
« less
Small-disturbance input-to-state stability of perturbed gradient flows: Applications to LQR problem
This paper studies the effect of perturbations on the gradient flow of a general nonlinear programming problem, where the perturbation may arise from inaccurate gradient estimation in the setting of data-driven optimization. Under suitable conditions on the objective function, the perturbed gradient flow is shown to be small-disturbance input-to-state stable (ISS), which implies that, in the presence of a small-enough perturbation, the trajectories of the perturbed gradient flow must eventually enter a small neighborhood of the optimum. This work was motivated by the question of robustness of direct methods for the linear quadratic regulator problem, and specifically the analysis of the effect of perturbations caused by gradient estimation or round-off errors in policy optimization. We show small-disturbance ISS for three of the most common optimization algorithms: standard gradient flow, natural gradient flow, and Newton gradient flow.
more »
« less
- PAR ID:
- 10513252
- Publisher / Repository:
- Elsevier
- Date Published:
- Journal Name:
- Systems & Control Letters
- Volume:
- 188
- Issue:
- C
- ISSN:
- 0167-6911
- Page Range / eLocation ID:
- 105804
- Subject(s) / Keyword(s):
- Input-to-state stability (ISS) Learning-based control Policy optimization Gradient systems Linear quadratic regulator (LQR)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This paper examines the problem of real-time optimization of networked systems and develops online algorithms that steer the system towards the optimal trajectory without explicit knowledge of the system model. The problem is modeled as a dynamic optimization problem with time-varying performance objectives and engineering constraints. The design of the algorithms leverages the online zero-order primal-dual projected-gradient method. In particular, the primal step that involves the gradient of the objective function (and hence requires a networked systems model) is replaced by its zero-order approximation with two function evaluations using a deterministic perturbation signal. The evaluations are performed using the measurements of the system output, hence giving rise to a feedback interconnection, with the optimization algorithm serving as a feedback controller. The paper provides some insights on the stability and tracking properties of this interconnection. Finally, the paper applies this methodology to a real-time optimal power flow problem in power systems, and shows its efficacy on the IEEE 37-node distribution test feeder for reference power tracking and voltage regulation.more » « less
-
This paper focuses on a newly challenging setting in hard-label adversarial attacks on text data by taking the budget information into account. Although existing approaches can successfully generate adversarial examples in the hard-label setting, they follow an ideal assumption that the victim model does not restrict the number of queries. However, in real-world applications the query budget is usually tight or limited. Moreover, existing hard-label adversarial attack techniques use the genetic algorithm to optimize discrete text data by maintaining a number of adversarial candidates during optimization, which can lead to the problem of generating low-quality adversarial examples in the tight-budget setting. To solve this problem, in this paper, we propose a new method named TextHoaxer by formulating the budgeted hard-label adversarial attack task on text data as a gradient-based optimization problem of perturbation matrix in the continuous word embedding space. Compared with the genetic algorithm-based optimization, our solution only uses a single initialized adversarial example as the adversarial candidate for optimization, which significantly reduces the number of queries. The optimization is guided by a new objective function consisting of three terms, i.e., semantic similarity term, pair-wise perturbation constraint, and sparsity constraint. Semantic similarity term and pair-wise perturbation constraint can ensure the high semantic similarity of adversarial examples from both comprehensive text-level and individual word-level, while the sparsity constraint explicitly restricts the number of perturbed words, which is also helpful for enhancing the quality of generated text. We conduct extensive experiments on eight text datasets against three representative natural language models, and experimental results show that TextHoaxer can generate high-quality adversarial examples with higher semantic similarity and lower perturbation rate under the tight-budget setting.more » « less
-
Abstract In the slow bond problem the rate of a single edge in the Totally Asymmetric Simple Exclusion Process (TASEP) is reduced from 1 to for some small . Janowsky and Lebowitz posed the well‐known question of whether such very small perturbations could affect the macroscopic current. Different groups of physicists, using a range of heuristics and numerical simulations reached opposing conclusions on whether the critical value of is 0. This was ultimately resolved rigorously in Basu‐Sidoravicius‐Sly which established that . Here we study the effect of the current as tends to 0 and in doing so explain why it was so challenging to predict on the basis of numerical simulations. In particular we show that the current has an infinite order phase transition at 0, with the effect of the perturbation tending to 0 faster than any polynomial. Our proof focuses on the Last Passage Percolation formulation of TASEP where a slow bond corresponds to reinforcing the diagonal. We give a multiscale analysis to show that when is small the effect of reinforcement remains small compared to the difference between optimal and near optimal geodesics. Since geodesics can be perturbed on many different scales, we inductively bound the tails of the effect of reinforcement by controlling the number of near optimal geodesics and giving new tail estimates for the local time of (near) geodesics along the diagonal.more » « less
-
Abstract We introduce a novel geometry-informed irreversible perturbation that accelerates convergence of the Langevin algorithm for Bayesian computation. It is well documented that there exist perturbations to the Langevin dynamics that preserve its invariant measure while accelerating its convergence. Irreversible perturbations and reversible perturbations (such as Riemannian manifold Langevin dynamics (RMLD)) have separately been shown to improve the performance of Langevin samplers. We consider these two perturbations simultaneously by presenting a novel form of irreversible perturbation for RMLD that is informed by the underlying geometry. Through numerical examples, we show that this new irreversible perturbation can improve estimation performance over irreversible perturbations that do not take the geometry into account. Moreover we demonstrate that irreversible perturbations generally can be implemented in conjunction with the stochastic gradient version of the Langevin algorithm. Lastly, while continuous-time irreversible perturbations cannot impair the performance of a Langevin estimator, the situation can sometimes be more complicated when discretization is considered. To this end, we describe a discrete-time example in which irreversibility increases both the bias and variance of the resulting estimator.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1016/j.sysconle.2024.105804
