Ambient Lighting Controller Based on Reinforcement Learning Components of Multi-Agents

Abstract

The paper presents a vision of sustainable eco-social laboratory, the ESLab which might be used to speed up the process of development of the recently proposed by authors of the Smart Eco-Social Apartment.  It is presented the multi-agent model of the ambient comfort measurement and environment control system to be used for the development of the ESLab.  The human Ambient Lighting Affect Reward index, the ALAR index is proposed at the first time used for development of the Reinforcement Learning Based Ambient Comfort Controller, the RLBACC for the ESLab. The ALAR index is dependent on human physiological parameters: the temperature, the ECG- electrocardiogram and the EDA-electro-dermal activity. The fuzzy logic is used to approximate the ALAR index function by defining two fuzzy inference systems: the Arousal-Valence System, and the Ambient Lighting Affect Reward (ALAR) System. The goal of the RLBACC is to find such the environmental state characteristics that create an optimal comfort for people affected by this environment. The Radial Basis Neural Network is used as the main component of the RLBACC to performing of two roles - the policy structure, known as the Actor, used to select actions, and the estimated value function, known as the Critic that criticizes the actions made by the Actor. The Critic in this paper was used as a value function approximation of the continuous learning tasks of the RLBACC. Ill. 9, bibl. 7 (in English; abstracts in English and Lithuanian).

DOI: http://dx.doi.org/10.5755/j01.eee.121.5.1656