Neuroscience Inspired Empathic Sensor System for Individual Comfort Climate Control
K. Kohlhof+, C. Messtorff*
+TH Cologne - University of Applied Sciences, Germany
*FOM Hamburg - University of Applied Sciences, Germany
Automatization processes are common within industry, health care, home application as well as consumer goods. They are meant to enable, improve or economize a work flow and thus assist men to facilitate their life. State-of-the-art controllers range from simple electro-mechanical to electronic PID-regulators, lookup tables, fuzzy logics and embedded systems. So, more and more complex control tasks may be realized, but mainly to process technical-physically sensor-signals like temperature, position or gas composition, e.g. in coffee machines, motion control or in combustion engines, respectively. But processes in relation to humans also need to consider their individual ways of perception of signals, i.e. to be empathic with their personal circumstances. So, air conditioning of rooms requires not only to control the environment surrounding men by measuring physically parameters like temperature, humidity or airflow. Also, the felt impact to men needs to be respected. Nowadays individual stress and health levels can non-invasively be derived from optical heart rate variability measurements. But other non-physically items including individual bio-, psycho- and social-statuses like age, gender, geographic or cultural origin are hardly to be accessed by sensors. A standard method in psychology to capture such personal data are questionnaires filled out by ticking boxes in dedicated item-scales, traditionally by paper-pencil and more modern by apps.
To consider such classified non-physically individual impacts together with technical-physically parameters both types of data recording is required: by sensors and also by abilities to digitally import non-physically items. For technological processing of such mixed data it is suggested to be inspired by neuroscientific principles optimized up to perfection in the long course of nature’s evolution process, e.g. by neurological signal processing within human brains. So, in analogy to human intelligence artificial intelligence is best suited to handle both, technical-physically data as well as digital classified personal data.
In this presentation an adapted sensor system implemented into an artificial neural network is sketched to control a room climate perceived by its occupants, evaluated based on measured physically parameters as well as on digitally captured questionnaire data. Artificial recurrent neural networks that are constructed following natural examples, e.g. for image processing or long and short termed memories, promise to regulate a room climate individually and empathically to people and to forecast its development for dynamic adaption.
My Profile
Introduction to Time Series Analysis with Ordinal Patterns
Ordinal Patterns are a non-parametric transformation of subsequences. They were proposed by Bandt & Pompe in 2002 as a means to unveil the underlying dynamics that govern time series. Ordinal Patterns have several interesting properties, among them being resistant to outliers and invariant to strictly increasing transformations. With far more than 2000 citations, they are a powerful tool in analysing time series. This talk is an overview of Ordinal Patterns, ways of analysing them, applications, and future research directions.
My Profile