System identification of nonlinear state-space models

By Thomas Bo Schön, Adrian G. Wills, Brett Ninness. at - Automatisierungstechnik, 2011

623 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/e0f4e5675e6f066987353f40ed49366f766cd81d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Thomas Bo Schön
2. Fredrik Lindsten
3. Marc Peter Deisenroth
4. Simo Särkkä
5. Niklas Wahlstrom
6. Nicolas Chopin
7. Andreas Svensson
8. Arnaud Doucet

In Control and Systems Engineering

1. Thomas Bo Schön
2. Marc Peter Deisenroth
3. Fredrik Lindsten
4. Niklas Wahlstrom
5. Simo Särkkä
6. Feng Ding
7. Carl E. Rasmussen
8. Lennart Ljung

Citations in the corpus

• Learning deep dynamical models from image pixels (2014, methodology): “In recent years, sequential Monte Carlo (SMC) methods have received attention for identifying nonlinear state-space models (Schön et al., 2011), see also the recent survey (Kantas et al.”
• Hierarchical multi-innovation stochastic gradient algorithm for Hammerstein nonlinear system modeling (2013, background): “Parameter estimation has wide applications in adaptive control, system identification [1–9], and filtering [10,11].”
• Identification of jump Markov linear models using particle filters (2014, methodology): “5 standard deviation (fields) H2 error for 7 runs of our RB-PSAEM using N = 3 particles (black) PSAEM [6] using N = 20 particles (blue) and PSEM [20] using N = 100 particles and M = 20 backward trajectories (red).”
• FedLoc: Federated Learning Framework for Data-Driven Cooperative Localization and Location Data Processing (2020, background): “Traditional SSM restricts both the transition function f(x) and the measurement function g(x) to empirical, parametric functions [10], whose parameters can be learned through the expectation-maximization (EM) algorithm [97] or Markov chain Monte Carlo (MCMC) algorithm [98].”
• Bayesian Filtering and Smoothing (2013, background): “3, the corresponding expectations can be approximated as follows (Schön et al., 2011).”
• Integrated pre-processing for Bayesian nonlinear system identification with Gaussian processes (2013, background): “More sophisticated approaches combining smoothing and system identification for state space models have the potential to give better results if one is ready to accept their additional complexity [8], [9].”
• Identification of Hammerstein-Wiener models (2013, background): “In previous work [36, 44, 17], the utility of choosing X as the time history of the full state vector of the underlying dynamics has been established.”
• Nonlinear System Identification: A User-Oriented Road Map (2019, background): “Concrete system identification examples were there is significant process noise available are provided in for example [153], [159], [160].”
• An efficient stochastic approximation EM algorithm using conditional particle filters (2013, methodology): “For instance, particle smoothing (PS) has been used within the EM algorithm (PSEM) for challenging identification problems [5] [6] [2].”
• On Particle Methods for Parameter Estimation in State-Space Models (2014, methodology): “3 as well as the generalized twofilter smoother have been used [87], [26], [27], [85], [8].”