Real-time personalised energy saving recommendations

• DOI: 10.1109/ithings-greencom-cpscom-smartdata-cybermatics50389.2020.00072
• Publication Date: 2020-12-28T20:52:07Z
• Authors: Christos Sardianos; Christos Chronis; Iraklis Varlamis; George Dimitrakop...; Yassine Himeur; Abdullah Alsalemi; Faycal Bensaali; Abbes Amira
• Keywords: persuasiveness; personalised; recommendations; 11. Sustainability; 0202 electrical engineering, electronic engineering, information engineering; real-time; 02 engineering and technology; recommender systems; 7. Clean energy; energy efficiency

The increased consumption of energy worldwide has boosted the interest of people for energy-efficient solutions at every level of daily life, from goods production and transportation to the use of household and office appliances. This gave rise to monitoring applications that monitor the daily user interaction with the electrical and electronic appliances, detect unnecessary or extensive usage and recommend corrective actions. In this direction, this work presents the anatomy of the Consumer Engagement Towards Energy Saving Behavior by means of Exploiting Micro Moments and Mobile Recommendation Systems (EM)3 recommendation engine, which supports household and office users with real-time personalized recommendations for avoiding unnecessary energy consumption and reducing the overall household (or office) energy footprint. The recommendation engine is based on a set of sensors that monitor energy usage, room occupancy, and environmental conditions inside and outside the living space, and a set of actuators that allow the remote control of devices, (e.g. on and off actions, set to eco or standby mode, etc.). The innovating feature of this recommendation engine is that it puts the human in the loop of energy efficiency by recommending actions at the right moment, in real-time, with user approval and rejection options. In addition, it provides savings related facts in order to increase the persuasiveness of the recommendations. Initial results show that users respond positively to personalized recommendations and are further persuaded when specific types of facts are chosen.