Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

ESP32-baserat larm – batteri och energioptimeringsstudie

Kadem, Ali LU and Chahrour, Ali LU (2026) In CODEN:LUTEDX/TEIE EIEL05 20261
Division for Industrial Electrical Engineering and Automation
Abstract (Swedish)
Köksbränder orsakas framförallt av mänskliga misstag vid matlagning. Traditionella brandvarnare fungerar reaktivt eftersom de aktiveras först när rök detekteras, vilket fördröjer tidiga insatser. Detta examensarbete undersöker utvecklingen av en batteridriven IoT-enhet för tidig branddetektion i köksmiljöer baserad på infraröd sensorteknik och digital signalbehandling. Arbetets primära fokus ligger på att optimera systemets energiförbrukning för att möjliggöra långsiktig autonom drift utan att kompromissa med en tillförlitlig
realtidsövervakning.
En jämförande studie genomförs mellan mikrokontrollers inom ESP32-familjen, specifikt från tillverkaren XIAO. ESP32-C3, ESP32-C6 och ESP32-S3, med avseende på processorarkitektur, trådlösa... (More)
Köksbränder orsakas framförallt av mänskliga misstag vid matlagning. Traditionella brandvarnare fungerar reaktivt eftersom de aktiveras först när rök detekteras, vilket fördröjer tidiga insatser. Detta examensarbete undersöker utvecklingen av en batteridriven IoT-enhet för tidig branddetektion i köksmiljöer baserad på infraröd sensorteknik och digital signalbehandling. Arbetets primära fokus ligger på att optimera systemets energiförbrukning för att möjliggöra långsiktig autonom drift utan att kompromissa med en tillförlitlig
realtidsövervakning.
En jämförande studie genomförs mellan mikrokontrollers inom ESP32-familjen, specifikt från tillverkaren XIAO. ESP32-C3, ESP32-C6 och ESP32-S3, med avseende på processorarkitektur, trådlösa kommunikationsprotokoll och implementering av
energisparlägen. Metodiken omfattar utvärdering av utvecklingskort samt analytisk och empirisk mätning av exekveringstider under olika mjukvarukonfigurationer. På grund av hårdvarubegränsningar modellerades strömförbrukningen analytiskt utifrån komponenternas datablad snarare än att mätas empiriskt.
Resultaten visar att mönsterkortets specifika hårdvaruoptimering och perifera kretsdesign har en mer avgörande inverkan på strömförbrukningen i djupsömn än den underliggande processorarkitekturen i sig.
Det specifika utvecklingskortet XIAO ESP32-S3 uppvisade en lägre strömförbrukning i viloläge jämfört med XIAO ESP32-C3, trots att det underliggande RISC-V-chippet på C3-modellen teoretiskt är strömsnålast. Detta illustrerar att mönsterkortets specifika hårdvaruoptimering har en mer avgörande inverkan i praktiken. Genom att kombinera en strömsnål kortlayout med en stringent duty cycling-strategi presterar XIAO ESP32-C6 bäst totalt. Resultaten visar att en underhållsfri, autonom drifttid på mellan 25 dagar och drygt 3,8 år (upp till 1389 dagar), beroende på valt samplingsintervall, är teoretiskt möjlig på ett
standardiserat 1500 mAh Li-Po-batteri. Värdena utgör en teoretisk övre gräns; den praktiska livslängden förväntas bli kortare. (Less)
Abstract
Residential kitchen fires represent a major safety challenge which is frequently caused byhuman error during cooking. Conventional fire alarms are reactive systems that trigger only when smoke reaches the sensors, by which time the fire is often established.
This thesis investigates the development of a proactive, battery powered IoT device capable of early fire detection through infrared sensing and signal processing. The system aims to distinguish actual flames from normal cooking patterns.
The primary focus of this research is the evaluation and optimization of power consumption to ensure long term autonomous operation while maintaining critical real-time monitoring focus.
A comparative study is conducted on three microcontrollers... (More)
Residential kitchen fires represent a major safety challenge which is frequently caused byhuman error during cooking. Conventional fire alarms are reactive systems that trigger only when smoke reaches the sensors, by which time the fire is often established.
This thesis investigates the development of a proactive, battery powered IoT device capable of early fire detection through infrared sensing and signal processing. The system aims to distinguish actual flames from normal cooking patterns.
The primary focus of this research is the evaluation and optimization of power consumption to ensure long term autonomous operation while maintaining critical real-time monitoring focus.
A comparative study is conducted on three microcontrollers from the ESP32-family which include ESP32-C3, ESP32-C6 and ESP32-S3 made by XIAO. By analyzing differences in their processor architecture, wireless communication protocols and the implementation of power saving modes such as deep sleep could the most suitable platform be identified. The methodology includes the evaluation of microcontroller development boards and empirical measurement of execution times across various software configurations. Due to hardware constraints, the power consumption is analytically modeled based on componentspecifications rather than empirically measured.
The results indicate that the specific hardware layout and board-level optimization have a more critical impact on deep sleep currents than the underlying processor architecture itself.
For instance, the specific development board XIAO ESP32-S3 exhibits a lower sleep current compared to the XIAO ESP32-C3, despite the underlying RISC-V chip on the C3 theoretically being more energy efficient. This illustrates the critical impact of the board's peripheral hardware design.
By combining a low-power board layout with strict duty cycling, the XIAO ESP32-C6
achieves the superior overall performance. The study indicates a theoretical maintenance-free operational time ranging from 25 days up to approximately 3.8 years on a standard 1500 mAh

Li-Po battery, depending on the chosen sampling interval. These figures represent a theoretical upper bound; the practical lifetime is expected to be shorter. (Less)
Please use this url to cite or link to this publication:
author
Kadem, Ali LU and Chahrour, Ali LU
supervisor
organization
course
EIEL05 20261
year
type
M2 - Bachelor Degree
subject
keywords
Internet of Things (IoT), Fire Detection, ESP32, Power Consumption, Fast Fourier Transform (FFT), Infrared Sensing, Deep Sleep, Duty Cycling.
publication/series
CODEN:LUTEDX/TEIE
report number
3152
language
Swedish
id
9245809
date added to LUP
2026-09-28 10:38:28
date last changed
2026-09-28 10:38:28
@misc{9245809,
  abstract     = {{Residential kitchen fires represent a major safety challenge which is frequently caused byhuman error during cooking. Conventional fire alarms are reactive systems that trigger only when smoke reaches the sensors, by which time the fire is often established.
This thesis investigates the development of a proactive, battery powered IoT device capable of early fire detection through infrared sensing and signal processing. The system aims to distinguish actual flames from normal cooking patterns.
The primary focus of this research is the evaluation and optimization of power consumption to ensure long term autonomous operation while maintaining critical real-time monitoring focus.
A comparative study is conducted on three microcontrollers from the ESP32-family which include ESP32-C3, ESP32-C6 and ESP32-S3 made by XIAO. By analyzing differences in their processor architecture, wireless communication protocols and the implementation of power saving modes such as deep sleep could the most suitable platform be identified. The methodology includes the evaluation of microcontroller development boards and empirical measurement of execution times across various software configurations. Due to hardware constraints, the power consumption is analytically modeled based on componentspecifications rather than empirically measured.
The results indicate that the specific hardware layout and board-level optimization have a more critical impact on deep sleep currents than the underlying processor architecture itself.
For instance, the specific development board XIAO ESP32-S3 exhibits a lower sleep current compared to the XIAO ESP32-C3, despite the underlying RISC-V chip on the C3 theoretically being more energy efficient. This illustrates the critical impact of the board's peripheral hardware design.
By combining a low-power board layout with strict duty cycling, the XIAO ESP32-C6
achieves the superior overall performance. The study indicates a theoretical maintenance-free operational time ranging from 25 days up to approximately 3.8 years on a standard 1500 mAh

Li-Po battery, depending on the chosen sampling interval. These figures represent a theoretical upper bound; the practical lifetime is expected to be shorter.}},
  author       = {{Kadem, Ali and Chahrour, Ali}},
  language     = {{swe}},
  note         = {{Student Paper}},
  series       = {{CODEN:LUTEDX/TEIE}},
  title        = {{ESP32-baserat larm – batteri och energioptimeringsstudie}},
  year         = {{2026}},
}