Mobiilsetele kasutajatele sensoritelt kogutud keskonnapõhiste andmete energiateadlik edastamine

Abstract

Tänapäeval levib järjest rohkem rakendusi, mis tajuvad ümbritsevat keskkonda ning pakuvad sellele põhinevalt kasutajale lisavõimalusi. Selliste võimaluste pakkumiseks on arendatud prototüüp, mis kogub keskkonna kohta andmeid kasutades Arduino platvormil põhinevad sensorite moodulit ning keskset andmet kogumise serverit.

Käesolevas töös arendati antud prototüüpi edasi, et tõsta aku vastupidavust ning seeläbi parandada lahenduse kasutatavust. Selleks loodi varieeruva sensoriandmete saatmise intervalliga lahendus, mis koosneb hägusloogikat kasutavast kontrollsüsteemist ning lihtsa lineaarse regressiooni mudelist. Lisaks loodi lahendus, mis lubab sensorite moodulil vabadel hetkedel minna puhkerežiimi.

Töö käigus asendati seni kasutuselolev XMPP protokoll HTTP protokolli vastu, et parandada ühenduse loomise ajakulu ning lubada sensorite moodulil kauem puhkerežiimis olla.

Parandatud lahenduse tulemusi mõõdeti mitme testi käigus. Kaks põhilist testi, mille käigus sensorite moodul sai voolu 9-voldiselt patareilt, andsid vastavalt tulemusteks 80 ja 110 \% pikema eluea. Sellest tulenevalt saab eeldada, et pakutud puhkerežiimi ja muutuva intervalliga andmete kogumist kasutav lahendus parandab aku vastupidavust ning seega ka prototüübi kasulikkust.

Nowadays, mobile applications are becoming more context aware due to technological achievements which enable the applications to anticipate users’ intentions. This is achieved through using the device’s own micromechanical artifacts that can be used to perceive the environment. However, this is constrained to the hardware limitations of devices as not all devices provide the same options. Moreover, perceiving the environment strains the battery and therefore has its impact on devices' everyday usage.

To remedy this, a proposed solution has been made in the thesis “Context Sensor Data on Demand for Mobile Users Supported by XMPP” by Kaarel Hanson. The solution is to gather environmental data by specialized sensor modules and store it in a data server. Afterwards, devices can query the data from the server and thus gain access to information beyond the capabilities of their own hardware.

The solution uses XMPP for transporting sensor data from Arduino microcontrollers (sensor modules) to the cloud. Arduino provides low-cost hardware, while the cloud offers the reliable and high- availability means for storing and processing sensor data. However, the developed prototype shows that running on a 9V battery the microcontroller lasts for 101 minutes when using an Ethernet module and 161,5 minutes with a WiFi module. These results are not good enough for remote data collection with limited access as the maintenance cost would be too high when the batteries need to be replaced frequently.

This thesis proposes an optimisation for the system so that instead of reading and sending sensor data every 10 seconds, the cloud server would notify the controller when to start sending data and when to stop. This means implementing an algorithm for detecting similar sensor data readings and notifying the microcontroller of needed operations. With similar readings, the microcontroller could be put to an idle state to limit power consumption, which would prolong battery life.

The aim is to optimise the sensor reading process enough to prolong Arduino microcontroller’s battery life on a 9V battery.