On developing theory of reservoir computing for sensing applications: the state weaving environment echo tracker (SWEET) algorithm
Artikel i vetenskaplig tidskrift, 2018

As a paradigm of computation, reservoir computing has gained an enormous momentum. In principle, any sufficiently complex dynamical system equipped with a readout layer can be used for any computation. This can be achieved by only adjusting the readout layer. Owning to this inherent flexibility of implementation, new applications of reservoir computing are being reported at a constant rate. However, relatively few studies focus on sensing, and in the ones that do, the reservoir is often exploited in a somewhat passive manner. The reservoir is used to post-process the signal from sensing elements that are placed separately, and the reservoir could be replaced by other information processing system without loss of functionality of the sensor ('reservoir computing and sensing'). An entirely different novel class of sensing approaches is being suggested, to be referred to as 'reservoir computing for sensing', where the reservoir plays a central role. In the State Weaving Environment Echo Tracker (SWEET) sensing approach, the reservoir functions as the sensing element if the dynamical states of the reservoir and the environment one wishes to analyze are strongly interwoven. Some distinct characteristics of reservoir computing (in particular the separability and the echo state properties) are carefully exploited to achieve sensing functionality. The SWEET approach is formulated both as a generic device setup, and as an abstract mathematical algorithm. This algorithmic template could be used to develop a theory (or a class of theories) of 'reservoir computing for sensing', which could provide guidelines for engineering novel sensing applications. It could also provide ideas for a creative recycling of the existing sensing solutions. For example, the Horizon 2020 project RECORD-IT (Reservoir Computing with Real-time Data for future IT) exploits the SWEET sensing algorithm for ion detection. Accordingly, the terms SWEET sensing algorithm and the RECORD-IT sensing algorithm can be used interchangeably.

sensing

the SWEET sensing setup

Reservoir computing

algorithmic template

Författare

Zoran Konkoli

Chalmers, Mikroteknologi och nanovetenskap (MC2), Elektronikmaterial och system

International Journal of Parallel, Emergent and Distributed Systems

1744-5760 (ISSN) 1744-5779 (eISSN)

Vol. 33 2 121-143

Ämneskategorier

Datorteknik

Reglerteknik

Datavetenskap (datalogi)

DOI

10.1080/17445760.2016.1241880

Mer information

Senast uppdaterat

2018-09-06