Patrolling Human Detection Robot Using Arduino and ESP Camera

Vol.4, No.1

Abstract

Patrolling robots have lately been used in the global industry as surveillance. The patrol robot
was usually used in large areas such as factories and subdivisions. This research develops an
autonomous robot using Arduino and ESP cameras that patrols, classifies, and notifies using
Arduino and ESP cameras. The robot was named CLAIRE, from the idea of Computer Vision
and Convolutional Neural Networks.

The result of this study shows that the diagrams, flowchart, and pin connection accomplished
their function. The robot patrols after the ESP32 Camera detects the lines; if it defines the
centroid to the left, the robot will turn to the left; if the centroid is defined to the right, then the
robot turns right and patrols until it reaches its destination. While on patrol, the camera rotates
vertically using servo motors. If the robot detects the students, the robot captures the FOV and
transmits a message to the user’s Telegram account with the captured images and their
classification. The result shows that the message was received by the user within 2 seconds
to 5 seconds upon detection. The respondents' perceptions in the study were gathered
through a prepared survey questionnaire that was separated into two parts.

The components used, such as the L293D Motor Driver, Ultrasonic Sensors, and DC Motors,
are compatible with the microcontroller, Arduino Uno R3, which shows success in patrolling.
The classification is processed by the RPI3B, and with the Internet, transmitting messages is a
success. Data gathered shows that the majority of the respondents are students aged 18-22,
and the majority of the respondents are also males and students. The researchers concluded
that Arduino and ESP 32 Camera are effective microcontrollers to help the patrolling human
detection robot navigate automatically. It helped to accomplish Patrolling Human Detection’s
main function to patrol, classify, and notify.