Portfolio

Active sensing

 

Through the years we have worked on several aspects of machine olfaction, ranging from instrumentation for solid-state and optical sensors to neuromorphic models of the olfactory pathway. Our current focus is on the development of active sensing strategies for tunable chemical sensors, that is, sensors whose selectivity towards different chemical species can be fine-tuned programmatically; this includes metal-oxide chemical sensors under temperature modulation and Fabry-Perot infrared interferometers.

The figure illustrates the process of active classification with an array of four metal-oxide sensors, ten temperatures per sensor, and a discrimination problem with six chemicals. At time t = 0, no information is available except that classes are a priori equiprobable: p(ωi) = 1/6. On the basis of this information, the active-classification algorithm decides to take the first sensing action (a1, measure sensor S2 at temperature T4), which leads to observation o1 and an updated posterior probability, p(ωi|o1, a1). After four sensing actions, evidence accumulated in the posterior p(ωi|o1, . . ., o4, a1, . . ., a4) and the cost of additional measurements are sufficient for the algorithm to assign the unknown sample to class ω3. In this toy example, an accurate classification is reached via only 10% of all sensor configurations.

We are also developing active-sensing strategies for Fabry-Perot Interferometers (FPI), devices that can act as tunable infrared (IR) spectrometers. IR spectroscopy provides a wealth of information to help estimate the identity and concentrations of chemicals, and allows us to analyse complex chemical mixtures without the need for physically separating them (i.e., via chromatography).

The video below illustrates the process of recognizing a chemical out of eight possible targets. The active sensing algorithm can identify the target by selecting only a few measurements, as opposed to having to capture the entire spectrum.

 

Relevant publications

J. Huang. R. Gutierrez-Osuna

Active wavelength selection for mixture identification with tunable mid-infrared detectors (Article)

Analytica Chimica Acta, in press, 2016.

(Links | BibTeX)

J. Huang, R. Gutierrez-Osuna

Detection of weak chemicals in strong backgrounds with a tunable infrared sensor (Inproceeding)

International Symposium on Olfaction and Electronic Nose, 2015.

(BibTeX)

J. Huang, R. Gutierrez-Osuna

Active wavelength selection for mixture analysis with tunable infrared detectors (Article)

Sensors and Actuators B: Chemical, 208, Page(s): 245–257, 2015.

(Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Active classification with arrays of tunable chemical sensors (Article)

Chemometrics and Intelligent Laboratory Systems, 132, Page(s): 91-102, 2014.

(Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Active temperature modulation of metal-oxide sensors for quantitative analysis of gas mixtures (Article)

Sensors and Actuators B: Chemical, 185, Page(s): 201-210, 2013.

(Links | BibTeX)

J. Huang, R. Gutierrez-Osuna

Active analysis of chemical mixtures with multi-modal sparse non-negative least squares (Conference)

38th International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2013.

(Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Active Decomposition and Sensing in Networks of Distributed Chemical Sensors (Techreport)

2012.

(Abstract | Links | BibTeX)

J. Huang, R. Gutierrez-Osuna

Active Analysis of Chemical Mixtures with Multi-modal Sparse Non-negative Least Sqares (Techreport)

2012.

(Abstract | Links | BibTeX)

J. Huang, R. Gosangi, R. Gutierrez-Osuna

Active Concentration-Independent Chemical Identification with a Tunable Infrared Sensor (Article)

Sensors Journal, IEEE, 2012.

(Abstract | Links | BibTeX)

J. Huang, R. Gosangi, R. Gutierrez-Osuna

Active Sensing with Fabry-Perot Infrared Interferometers (Conference)

Proceedings of the 14th International Symposium on Olfaction and Electronic Nose, 2011.

(Abstract | Links | BibTeX)

R. Gutierrez-Osuna, R. Gosangi, A. Hierlemann

Invited: Advances in Active and Adaptive Chemical Sensing (Conference)

Proceedings of the 14th International Symposium on Olfaction and Electronic Nose, 2011.

(Abstract | Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Quantification of Gas Mixtures with Active Recursive Estimation (Conference)

Proceedings of the 14th International Symposium on Olfaction and Electronic Nose, 2011.

(Abstract | Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Data-driven Modeling of Metal-oxide Sensors with Dynamic Bayesian Networks (Conference)

Proceedings of the 14th International Symposium on Olfaction and Electronic Nose, 2011.

(Abstract | Links | BibTeX)

R. Gutierrez-Osuna, A. Hierlemann

Adaptive Microsensor Systems (Article)

Annual Review of Analytical Chemistry, 3, Page(s): 255–276, 2010.

(Abstract | Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Active temperature programming for metal-oxide chemoresistors (Article)

Sensors Journal, IEEE, 10, 6, Page(s): 1075–1082, 2010.

(Abstract | Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Energy-aware active chemical sensing (Conference)

Proceedings of IEEE Sensors, 2010.

(Abstract | Links | BibTeX)

R. Gosangi, R. Gutierrez-Osuna

Active chemical sensing with partially observable Markov decision processes (Conference)

Proceedings of 13th International Symposium on Olfaction and Electronic Noses, 2009.

(Abstract | Links | BibTeX)

A. Hierlemann, R. Gutierrez-Osuna

Higher-order chemical sensing (Article)

Chemical reviews, 108, 2, Page(s): 563, 2008.

(Links | BibTeX)