Special Issue on Finite-time estimation, diagnosis and synchronization of uncertain systems
摘要截稿:
全文截稿: 2019-03-01
影响因子: 1.54
期刊难度:
CCF分类: 无
中科院JCR分区:
• 大类 : 计算机科学 - 4区
• 小类 : 自动化与控制系统 - 4区
Overview
The estimation problem of unmeasured states, unknown parameters or exogenous inputs for complex dynamics is a prerequisite for the implementation of a control strategy in practice. Emerging new fields of applications, such as biology, renewable energies, Internet of things, etc., necessitate reliable estimation algorithms for diverse classes of uncertain systems admitting new restrictions and requirements. Conventional approaches for estimation are based on the paradigm of asymptotic statistics and convergence, while in many areas the life horizon of a system is finite, and an action has to be performed on a finite interval of time as well. For instance, rapid case detection and response are key to ending a disease outbreak through efficient estimation of its parameter and subsequent decision making. Indeed, the efficacy of healthcare interventions is known to be higher during the burst stage of the epidemic episode. Another example: in a smart grid, after its commutation or a short circuit, the supervision system has only few seconds to take a decision about the new configuration, next guarding electronics may switch off the system for safety, which also requires fault detection and isolation tools working in finite time.
To ensure a non-asymptotic estimation (and control), many convergence concepts have been proposed: finite-time stability, fixed-time stability, hyper-exponential stability, stability in prescribed time. All these notions allow the design of control and estimation algorithms converging in a finite interval of time, then answering the requirements of new developing fields of application. To achieve convergence in finite or fixed-time, most approaches resort to high-gain injection, which calls for particular attention when deployed in noisy environments. Therefore, it is of primary importanceto characterize the robustness of this class of algorithms with respect to measurement noise, sensor delays and exogenous perturbations.
This special issue aims at presenting the last innovative results for design methods of finite-time estimation algorithms and their applications to fault detection and isolation or synchronization. An accompanying objective consists in the stimulation of a discussion on this subject and a comparison of existing finite-time and asymptotical approaches.
The special issue is devoted to all topics related to finitetime estimation and its applications, including (but not limited to) the following subjects:
- Finite/fixed-time (hyper-exponential or in prescribe time) algorithms for estimation of state, parameters or inputs;
- Parameter estimation of multi-sinusoidal signals and exponentially fading sinusoids;
- Benchmarking and comparison of finite-time and asymptotically estimating tools;
- Fault detection and isolation in a finite time Finite-time synchronization;
- Robustness analysis of non-asymptotic estimation algorithms;
- Discrete-time realizations of finite-time estimation algorithms;
- Tuning rules for convergence rate and robustness improvement;
- Applications for emerging and conventional areas of science and technology.