Simulacija realnega senzorja preko pythonskega vmesnika za navidezno pretakanje podatkov v Arm virtualni strojni opremi

MAHNIČ, TEO (author), Bürmen, Arpad (mentor)

V magistrski nalogi je predstavljena razširitev pythonovskega vmesnika za navidezno pretakanje podatkov (VSI), ki v povezavi z Arm virtualno strojno opremo (AVH) omogoča natančno simulacijo realnega senzorja. Osnovni pythonov VSI modul je razširjen z medpomnilnikom FIFO in skupino funkcij, ki omogočajo procesiranje podatkovnih datotek SDS. Slednje vsebujejo surove podatke senzorja in časovne žige, pri katerih so bili podatki zajeti. Časovni žigi omogočajo, da se podatki aplikaciji strežejo v enakih intervalih, kot če bi prihajali neposredno iz senzorja, s katerim so bili zajeti. V ta namen VSI računa drseče povprečje intervalov med podatki v posnetku SDS znotraj nastavljenega časovnega okna. Izračunani intervali se nato uporabijo za nastavljanje časovnika, ki proži prekinitveno rutino v aplikaciji. Končni uporabnik lahko prilagaja interval drsenja in velikost časovnega okna glede na hitrost simuliranega senzorja, s čimer se približa delovanju realnega senzorja.

• Language: Slovenian
• Keywords: Arm virtualna strojna oprema, navidezni vmesnik za pretakanje, strojno učenje, simulacija, senzor, SDS posnetek
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2023
• PID: 20.500.12556/RUL-151589
• COBISS.SI-ID: 169885443
• Publication date in RUL: 11.10.2023

Secondary language

• Language: English
• Title: Simulating a real-world sensor via the Python interface for virtual data transmission in Arm virtual hardware

Abstract

The master's thesis presents an extension of the Python interface for virtual data transmission (VSI), which in conjunction with Arm virtual hardware (AVH) enables an accurate simulation of a real sensor. The basic Python VSI module is extended with a FIFO buffer and a set of functions that enable the processing of SDS data files. The files contain raw sensor data and the corresponding time stamps at which the sensor data was captured. Timestamps allow the data to be served to the application at the same intervals as if it were coming directly from the sensor with which it was captured. For this purpose VSI calculates a sliding average of the intervals between the data in the SDS recording within the set time window. The calculated intervals are then used to set a timer that triggers an interrupt routine in the application. The end user can adjust the sliding interval and the size of the time window according to the speed of the simulated sensor thus closely reproducing the performance of the real sensor.

• Keywords: Arm virtual hardware, virtual streaming interface, machine learning, simulation, sensor, SDS record