Cambot

Abstract

The increasing need for hands-free video recording in dynamic environments has motivated the development of robotic camera platforms. This project presents Cambot, a prototype mobile robot designed to autonomously track and film a moving subject using onboard image processing while also supporting manual control via a wireless controller. The system architecture integrates multiple microcontrollers for distributed task management: an Arduino Nano for controller input and NRF24 communication, an Arduino Uno for decoding NRF24 signals, and an Arduino Mega for motor control, ultrasonic obstacle avoidance, and LCD display output. A Raspberry Pi 5 with an OAK-D camera performs real- time image recognition and sends movement commands to the Mega. Mobility is achieved using four Omni-wheels driven by dual L298N motor drivers, while a 12 V/9 A battery with XL4015 buck converters ensures stable power distribution. Key features include autonomous person-tracking, manual override through NRF24 wireless control, and security access using an RFID system. Safety is enhanced by ultrasonic sensors on all four sides that halt movement if an obstacle is detected within 20 cm. Power management is improved by splitting loads into two switchable subsystems—one dedicated to the Raspberry Pi + OAK-D, and the other to the motor control system. System performance was evaluated through video testing, power consumption measurements, and real-time LCD feedback. The project demonstrates a scalable and energy-aware approach to robotic filming, providing a prototype suitable for hobbyist and semi-professional use in mobile videography.