Page 31 - AJWEP-v22i2

Robotic pool cleaning for better hygiene

Figure 1. Swimming pool robot functional units
Abbreviations: BLDC: Brushless direct current motor; DC: Direct current; GPS: Global Positioning System.

quality monitoring unit. The power unit is essential surface cleaning equipment are packed inside the
for the robot’s functionality, as it must operate both robot cleaning machine. In this case, the Pixy CMU
inside and outside the swimming pool. Therefore, a Cam5 vision sensor is utilized. This sensor features
water-resistant power unit is selected. Solid-state and a dual-core ARM Cortex-A72 processor capable of
lithium-ion batteries are recommended due to their operating at a frequency of 1.5 GHz, which is widely
waterproofing capabilities. Regardless of the battery used in DIY robotics, home automation, and research
type, it is crucial to ensure that it is adequately sealed projects involving image processing. It also includes
or housed in a waterproof compartment to prevent a unidirectional image sensor, 128k RAM, 2 MB flash
damage attributed to water exposure. In addition, memory, and various communication interfaces such as
weight considerations are important; for smaller CAN, RS-232, UART, SPI, I2C, and USB. The vision
aquatic vehicles, lightweight options such as lithium- sensor employs color moments because this algorithm
24
ion or lithium-polymer batteries are prioritized. Finally, is useful in recognizing objects and classifying tasks to
implementing battery management systems is essential detect objects, specifically using a method known as the
to monitor battery health and prevent overcharging, color-connected components algorithm. This algorithm
especially for lithium-based batteries. was implemented on the ARM Cortex-A72 processor
The sensor unit has two kinds of sensors: a GPS using JAVA. The color moments algorithm operates with
sensor that defines the boundary of the swimming high speed, efficiency, and reliability. In this context,
pool and a vision sensor that identifies the objects in statistical measures of color distribution were used in
the pool. A high-resolution vision sensor and water an image for filtering. The algorithm calculates the first

Volume 22 Issue 2 (2025) 25 doi: 10.36922/ajwep.6564