Zaznavanje pitja na senzorski zapestnici s pomočjo strojnega učenja

CERGOLJ, VINCENT (author), Pirc, Matija (mentor), Luštrek, Mitja (co-mentor)

Kljub pomembnosti vzdrževanja ustreznega stanja hidracije človeškega telesa je vnos vode zaradi hitrega tempa življenja ljudi pogosto zanemarjen. Za starejše in ljudi s primanjkljaji na kognitivnem področju je lahko zmanjšan vnos vode še bolj skrb zbujajoč, ne le zaradi škodljivega vpliva dehidracije, ampak tudi zato, ker so mehanizmi za uravnavanje hidracije pri teh osebah običajno manj učinkoviti. Navade, povezane s pitjem, imajo ključno vlogo pri splošnem zdravju. S stalnim spremljanjem vnosa tekočine pridobimo informacije, ki so lahko izredno uporabne pri ukrepanju zoper nezdravih navad pitja in imajo pomembno vlogo pri preprečevanju številnih bolezni. Pričujoča diplomska naloga obravnava razvoj nove metode strojnega učenja za preiskovanje pitja v naravnih okoljih. Glavni namen te naloge je bil razvoj programa za zaznavanje pitja, ki deluje neposredno na mikrokrmilniku s čim manjšo porabo električne energije. Temelji na podatkih iz inercijskih senzorjev, vgrajenih v praktično, neinvazvino zapestno napravo, ki ves dan spremlja gibanje zapestja in samodejno zazna dogodke pitja. Za razvoj robustne metode smo zasnovali poseben postopek zbiranja podatkov, s katerim smo zbrali 135 ur podatkov, od tega 2 uri in 30 minut ustreza aktivnostim pitja. Za potrditev uspešnosti predlagane metode smo opravili obsežno vrednotenje na računalniku in neposredno na zapestnici. Izkazalo se je, da je najboljši model XGBoost z rezultatom natančnosti 94,3 % in priklica 83,5 % pri oknu dolžine 4,92 s in 50 % prekrivanjem. Izvedli smo preizkus v kontroliranem okolju z vnaprej določenimi aktivnostmi pitja in ne-pitja. Dosegli smo rezultat priklica 76,0 %. Nazadnje je nekaj prostovoljcev nosilo zapestnico čez dan v prosto življenjskih pogojih, kjer so bili rezultati natančnosti 74,5 % in priklica 89,9 %. Analiza energetske učinkovitosti je pokazala, da lahko ob zmernem proženju metode pričakujemo avtonomno delovanje zapestnice tudi do 203 dni. Doseženi rezultati kažejo, da je predlagana metoda sposobna zaznati dogodke pitja v scenariju iz vsakdanjega življenja z uporabo podatkov iz inercijskih senzorjev, nošenih na zapestju, in je dovolj robustna, da se obnese na podatkih udeležencev, o katerih ni imela predhodnega znanja.

• Language: Slovenian
• Keywords: prepoznavanje dejavnosti, senzorska zapestnica, zaznavanje pitja, strojno učenje, klasifikacija, mikrokrmilnik, inercijski senzorji
• Work type: Bachelor thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2022
• PID: 20.500.12556/RUL-139132
• COBISS.SI-ID: 119875331
• Publication date in RUL: 31.08.2022

Secondary language

• Language: English
• Title: Drinking event detection on a sensing wristband using machine learning

Abstract

Despite the importance of maintaining an adequate hydration status, water intake is frequently neglected due to the fast pace of people's lives. For the elderly and people with cognitive deficits, reduced water intake can be even more worrying, not only because of the detrimental effects of dehydration, but also because the mechanisms to regulate hydration in these individuals are usually less effective. It is well established that drinking related habits play a critical role in overall human health. By constantly monitoring fluid intake, we gain information that can be extremely useful in dealing with unhealthy drinking habits and play an important role in preventing many diseases. This thesis deals with the problem of developing a novel machine learning method for drinking detection in naturalistic settings. The main focus was the development of a drinking detection program that works directly on the microcontroller with the lowest possible power consumption. It is based on data from inertial sensors built into a practical, non-invasive wrist-worn device that monitors wrist movement throughout the day and automatically detects drinking events. To develop robust methods, we designed a special data collection procedure, which collected 135 hours of data, of which 2 hours and 30 minutes corresponds to drinking activities. To validate the performance of the proposed method, an extensive evaluation was carried out on a computer and directly on the wristband. The best model turned out to be XGBoost with an accuracy of 94.3 % and a recall of 83.5 % with a signal window length of 4.92 s and 50% overlap. We also performed a test in a controlled environment with predefined drinking and non-drinking activities. We achieved a recall score of 76.0 %. Finally, some volunteers wore the bracelet during the day in free living conditions, with a precision of 74.5 % and recall of 89.9 %. Energy efficiency analysis showed that with moderate triggering of the method, we can expect autonomous operation of the wristband for up to 203 days. The achieved results suggest that the proposed method is capable of detecting drinking events in a free-living scenario using data from wrist-worn inertial sensors and is robust enough to cope with data from participants about whom it had no prior knowledge.

• Keywords: activity recognition, sensing wristband, drinking detection, machine learning, classification, microcontroller, inertial sensors