Journal
IEEE TRANSACTIONS ON AUTOMATIC CONTROL
Volume 61, Issue 1, Pages 182-187Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TAC.2015.2426291
Keywords
Nonlinear system identification; re-weighted l(1)-minimization; sparse Bayesian learning
Funding
- Microsoft Research
- EPSRC [EP/I03210X/1, EP/G036004/1, EP/M002187/1]
- FNR
- ERASysAPP
- Engineering and Physical Sciences Research Council [EP/G036004/1, EP/M002187/1, EP/K020617/1, EP/I03210X/1] Funding Source: researchfish
- EPSRC [EP/K020617/1] Funding Source: UKRI
Ask authors/readers for more resources
This technical note considers the identification of nonlinear discrete-time systems with additive process noise but without measurement noise. In particular, we propose a method and its associated algorithm to identify the system nonlinear functional forms and their associated parameters from a limited number of time-series data points. For this, we cast this identification problem as a sparse linear regression problem and take a Bayesian viewpoint to solve it. As such, this approach typically leads to nonconvex optimizations. We propose a convexification procedure relying on an efficient iterative re-weighted l(1)-minimization algorithm that uses general sparsity inducing priors on the parameters of the system and marginal likelihood maximisation. Using this approach, we also show how convex constraints on the parameters can be easily added to the proposed iterative re-weighted l(1)-minimization algorithm. In the supplementary material available online (arXiv: 1408.3549), we illustrate the effectiveness of the proposed identification method on two classical systems in biology and physics, namely, a genetic repressilator network and a large scale network of interconnected Kuramoto oscillators.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available