Digital Twin to Prevent Lonely Death
Introduction
In modern society, advances in medicine and personal hygiene have increased life expectancy, but they have also led to issues like indifference towards single-person households, resource wastage, and energy inefficiency. The rise of single-person households has resulted in problems such as social isolation and elderly loneliness. Additionally, the evolution of sensor technology, wireless networks, and AI is transforming society, with efforts to collect power consumption data for analyzing electricity usage patterns in buildings and promoting safety. Non-Intrusive Load Monitoring (NILM) technology is being advanced using neural networks, but obtaining real-world data presents challenges. To address this, the approach is to use artificially synthesized data from digital twin houses, which simulate daily activities and capture regional and individual characteristics for research purposes.
Configuring a Digital Twin Environment
we create virtual houses in 4 types and some types are based on reality, there can be more than 1 resident, so we create more virtual houses for style that correspond to number of people. Additionally, for the Avatar simulation that we did from a random daily activity from the Life Scenarios spreadsheet, we also used Unity in the same way as creating a virtual home. But because of randomness has various conditions such as the probability of using that type of electrical appliance, season, and timing. In order to define these conditions, we have to create a C# script to assign those conditions to each electrical appliance and for avatar movements and gestures to correspond with the type of electrical appliances.
Generation of Synthetic Data
when the simulation ends, we will get synthetic data. Therefore, we have created a C# script for visualizing synthetic data that we get by calculating the total power consumption and total used time for every electrical appliance in each type of house. The principle calculated total power consumption and total used time for each electrical appliance by taking the results from synthetic data that collected the duration of use of that electrical appliance at each time and combining them to get total used time. Then, use that total used time multiplied by 60 to convert minutes into seconds. Because of power consumption unit is watt, it`s 1 joule per second after converting used time from minute to second. Then, multiply with power consumption(watt) of that electrical appliance to get the total power consumption of that electrical appliance. After that all the result that already visualizing will be writing on to CSV format to make it easier to use and analyze the results further.
Papers
Dong Hyeon Kim, Rinrada Tirasirichai, Jun Hyeok Jang, Hyeon Hoo Hwang , Soon Ki Jung, Exploring Street view API connection to improve user experience in digital twin environment , 6th International Conference on Culture Technology(ICCT 2023), (2023.1.1 ~ 2023.1.4)
Jun Hyeok Jang, Rinrada Tirasirichai, Dong Hyeon Kim, Jin Ho Lee, Soon Ki Jung, Implementation of Digital Twins in Smart Homes and Generation of Remote Meter Data , 6th International Conference on Culture Technology(ICCT 2023), (2023.1.1 ~ 2023.1.4)
장준혁, 이진호, 허세환, 황현후, 김동현, Rinrada Tirasirichai, 정순기, 일상 생활 활동 추정을 위한 데이터셋 탐구, 2023 한국멀티미디어학회 추계학술발표대회(2023.11.17 ~ 2023.11.18)