← Selected projects

03 · Robotics and computer vision project

AI ArmCar

A robotics prototype bringing computer vision, embedded control, and browser-based vehicle interaction into one exploratory project.

Year
2023
Focus
Object detection · Computer vision integration · Embedded control · Hardware and software integration · Browser control interface · Cross-language integration
Stack
Python · C++ · Arduino · ESP32 · YOLOv5 · HTML
Protocols
WebSocket
Platform
ESP32
Implementation
YOLOv5 computer-vision components · ESP32 and Arduino sketches · Camera stream handling · Browser control interface · Motor and servo control
View repository

Overview

AI ArmCar brings together computer-vision software, embedded-control experiments, and browser interfaces for an arm-car prototype. The repository is exploratory, with multiple experiments around vehicle movement, sensors, camera handling, and an articulated arm.

Context

The project explores how a browser interface can communicate with ESP32-based control code while camera and computer-vision components support the prototype. Its repository includes both project-specific control files and a YOLOv5 codebase, so the portfolio distinguishes the integration work from claims about the underlying detection framework.

Implementation

The repository includes Python computer-vision components, C++ source files, ESP32 and Arduino sketches, and HTML control pages. One ESP32 camera sketch hosts a page with movement, speed, light, pan, and tilt controls; separate WebSocket channels carry camera data and control messages to the vehicle hardware.

Execution flow

Browser control → WebSocket control message → ESP32 sketch → motor, light, or servo behavior. Camera frames travel through a separate WebSocket channel to the browser interface.

Technical decisions

Separating camera and control WebSockets keeps streamed image data distinct from small control messages. The embedded interface also initializes control values when a browser connection opens, giving the hardware a defined starting state.

Evidence

The project repository contains the embedded HTML control page, ESP32 camera and motor code, Arduino sketches, Python files, and C++ experiments. It does not claim that every included YOLOv5 source file was authored for the prototype.