1,721,011 research outputs found
A Statistical Learning Theory Approach for the Analysis of the Trade-off Between Sample Size and Precision in Truncated Ordinary Least Squares
No full textThis chapter deals with linear regression problems for which one has the possibility of varying the supervision cost per example, by controlling the conditional variance of the output given the feature vector. For a fixed upper bound on the total available supervision cost, the trade-off between the number of training examples and their precision of supervision is investigated, using a nonasymptotic data-independent bound from the literature in statistical learning theory. This bound is related to the truncated output of the ordinary least squares regression algorithm. The results of the analysis are also compared theoretically with the ones obtained in a previous work, based on a large-sample approximation of the untruncated output of ordinary least squares. Advantages and disadvantages of the investigated approach are discussed
Hierarchical switching for active disturbance attenuation with fine controller tuning
No full textIn this paper, a novel adaptive disturbance attenuation algorithm is proposed combining switching and tuning. A two-level hierarchical switching logic is developed, which first selects in a short time the potentially best controller among a finite pre-designed family and then performs a local refinement of its attenuation capability. Thanks to the controller fine tuning, the proposed technique is able to provide a substantial performance improvement in terms of attenuation level as compared with a pure adaptive switching control scheme; at the same time, it retains the positive features of switching-based approaches, in particular, concerning the possibility of rapidly achieving a satisfactory behavior. Further, an arbitrary attenuation level is ensured in the presence of particular classes of disturbances and provided that it is compatible with robust stability requirement. Simulation results are shown to underline the potential of the approach as a solution to the problem. Copyright © 2016 John Wiley & Sons, Ltd
A distributed Kalman filter with event-triggered communication and guaranteed stability
No full textThe paper addresses Kalman filtering over a peer-to-peer sensor network with a careful eye towards data transmission scheduling for reduced communication bandwidth and, consequently, enhanced energy efficiency and prolonged network lifetime. A novel consensus Kalman filter algorithm with event-triggered communication is developed by enforcing each node to transmit its local information to the neighbors only when this is considered as particularly significant for estimation purposes, in the sense that it notably deviates from the information that can be predicted from the last transmitted one. Further, it is proved how the filter guarantees stability (mean-square boundedness of the estimation error in each node) under network connectivity and system collective observability. Finally, numerical simulations are provided to demonstrate practical effectiveness of the distributed filter for trading off estimation performance versus transmission rate
Distributed Kalman filtering with data-driven communication
No full textThe paper deals with distributed Kalman filtering over a peer-to-peer sensor network with focus on a data transmission scheduling strategy aiming at reduced communication bandwidth and, consequently, at enhanced energy efficiency and prolonged network lifetime. A novel distributed Kalman filter algorithm with data-driven communication is devised relying on the idea that each node transmit its local information to the neighbors only when this is deemed to be particularly relevant for estimation purposes, i.e. whenever it significantly deviates from the information predicted from the last transmitted one. An interesting information-theoretic interpretation of the proposed strategy is presented and numerical simulations are provided to demonstrate its practical effectivenes
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Adaptive disturbance attenuation via logic-based switching
No full textThe problem of attenuating unknown and possibly time-varying disturbances acting on a linear time-invariant dynamical system is addressed by means of an adaptive switching control approach. Given a family of pre-designed stabilizing controllers, a supervisory unit infers in real-time the potential behavior of each candidate controller and selects the one providing the best potential performance. To this aim, a set of test functionals is defined, which is shown to enjoy favorable inference properties under certain assumptions on the nature of the disturbances. Both persistent-memory and finite-memory test functionals are analyzed. Further, an implementation of the switching controller is proposed which always guarantees stability of the feedback loop, even if the disturbance characteristics are such that the switching is persistent. Simulation results are provided to show the effectiveness of the proposed method
Optimal direct data-driven control with stability guarantees
No full textFor model-free optimal control design, this paper proposes an approach based on optimizing the reference model that is used in direct data-driven controller synthesis. Optimality is defined with respect to suitable cost functions reflecting desired performance and control objectives. We rely on the well-known Virtual Reference Feedback Tuning technique and on a direct control design approach that ensures stability of the resulting closed-loop system. The proposed design method leads to a non-convex optimization problem with a small number of variables that can be easily solved by a global optimizer, such as by particle swarm optimization. The effectiveness of the proposed solution is illustrated in simulation examples
Principal component analysis applied to gradient fields in band gap optimization problems for metamaterials
No full textA promising technique for the spectral design of acoustic metamaterials is based on the formulation of suitable constrained nonlinear optimization problems. Unfortunately, the straightforward application of classical gradient-based iterative optimization algorithms to the numerical solution of such problems is typically highly demanding, due to the complexity of the underlying physical models. Nevertheless, supervised machine learning techniques can reduce such a computational effort, e.g., by replacing the original objective functions of such optimization problems with more-easily computable approximations. In this framework, the present article describes the application of a related unsupervised machine learning technique, namely, principal component analysis, to approximate the gradient of the objective function of a band gap optimization problem for an acoustic metamaterial, with the aim of making the successive application of a gradient-based iterative optimization algorithm faster. Numerical results show the effectiveness of the proposed method
Direct Control Design via Controller Unfalsification
No full textThis paper proposes a non-iterative direct approach for controller design from experimental data; the parameters of a controller of a prescribed order and structure are optimized with respect to a relevant performance criterion. The proposed design method enjoys the following features: (i) It does not involve the identification of the process to be controlled; (ii) it only requires a single experiment; (iii) in the case of stable plants, no initial controller is needed even when the process to be controlled is non-minimum phase; and (iv) it provides sufficient conditions for the resulting closed-loop system to be stable. The approach builds upon the so-called unfalsified control theory; this key point makes it possible to derive simple and intuitive relations between the choice of the performance criterion to be optimized and closed-loop stability conditions. The analysis is supported by numerical examples
- …
