Revived a shorted hoverboard inverter by troubleshooting voltage regulators and designing a low voltage cutoff system. Built a custom ESP8266 transmitter/receiver along with an android-based object-recognition AI mobile assistant.
This project demonstrates the creative reuse of recycled hardware to develop a motorized trolley that integrates wireless control and AI-driven capabilities. The system is built using components from an old drone controller and two damaged smart plugs.
Control of the trolley is handled through an Arduino-based system that reads joystick inputs from the drone controller while the ESP8266 radio module handles the wireless communication. A mode switch allows the user to alternate between wireless joystick control and an AI-assisted app. In AI mode, the app sends signals via USB to a second Arduino, enabling advanced functionality.
One standout feature is the AI-powered person detection system. A smartphone mounted on the trolley uses the app to recognize and track a person in front of the cart. Once a person is detected, the trolley autonomously follows them, adjusting its speed and direction to keep the person in view of the camera, ensuring continuous tracking.
To enhance cost efficiency, a low-pass filter was used to convert fast DC pulses into smooth analog signals, enabling reliable motor control and smooth operation of the trolley.
