Chess Download Center

ISEA Dataset

Welcome to the Ion-Selective Electrode Array (ISEA) dataset page. This page contains datasets acquired in experiments with arrays of ion-selective electrodes (ISE). These datasets may be helpful when developing supervised or unsupervised signal processing algorithms for quantitative analysis through smart ISE arrays.

The ISEA dataset is a fruit of a cooperation between the GIPSA-lab and the LAAS-CNRS.

Smart sensor arrays

Sensor array

Ion-selective electrodes (ISEs) provide a practical and cheap way for estimating the ionic activity (a measure of effective concentration of an ion). The most known example of ISE is the glass electrode used for measuring the pH value. Besides this ubiquitous example, there are ISEs designed for other ions that are relevant in applications such as water quality control, food industry, and blood analysis.

Despite all the advantages in the use of ISEs, this device usually lacks selectivity. For example, in a solution containing the ammonium and potassium ions, the response of an ammonium ISE may also depend on the concentration of the potassium ion and vice-versa.

A possible way to overcome this interference problem in an ISE is based on the idea of smart sensor arrays (SSA). In this approach, the measurements are conducted by an array of electrodes that are not necessarily selective. Then, the acquired data are treated by signal processing algorithms whose aim is to retrieve the relevant information, which can be either qualitative or quantitative.

The application of smart sensor arrays (SSA) on chemical sensing problems has been attracting a great deal of attention in the last years, notably in the design of electronic tongues and noses. By relying on methods developed in the domains related to the information processing (signal processing, statistics, machine learning, etc), SSAs provide a flexible way for conducting qualitative and quantitative chemical analysis.

The signal processing algorithm used in a SSA can be of two types. In a first approach, which is called supervised learning, a training dataset (calibration points) containing a set of input-output measurements is used, before the effective use of the sensor array, for adjusting the parameters of the signal processing algorithm. In a second approach, often denoted as unsupervised (or blind) learning, the data processing is conducted with any reference training points, i.e. the signal processing algorithm uses only the array response.

No matter the type of signal processing algorithm used in an ISE array, it is quite helpful to have access to training datasets when developing new algorithms. This is true even for blind methods. Motivated by that, the goal of this page is to make publicly available the results of a set of experiments with ISE arrays that were performed in the context of the development of blind signal processing algorithm for quantitative analysis. Note, however, that the acquired dataset can also be useful for researches interested in supervised methods for quantitative analysis.

Please fill the form before downloading the data. This will help us to communicate new experiments and keep you informed about our work. After filling the form, a link will be sent to the e-mail address entered.

Name: Mail: Affiliation: Country: