Welcome to the Robocar project! This repository contains the source code for an autonomous vehicle capable of line following and obstacle detection.

The project is organized into three main components:

This directory contains the core intelligence of the autonomous car. It uses Computer Vision and Deep Learning to navigate.

• AI & Computer Vision: Uses a TensorFlow Lite model (128_mask_gen.tflite) to segment the lane lines from the camera feed.
• Path Planning: Processes the segmented mask using raycasting to determine the optimal steering angle and speed.
• VESC Control: Interfaces with the VESC motor controller to drive the car.
• Dual Modes: Supports both autonomous driving (IA mode) and manual control via gamepad (CONTROLLER mode).

Usage:

# Run in autonomous mode
python3 main.py --mode ia

# Run in manual controller mode
python3 main.py --mode controller

A high-performance C++ utility for interfacing with the LD19 LiDAR sensor.

• Fast Parsing: Efficiently parses raw serial data from the LiDAR.
• Visualization: Includes a real-time visualization tool built with SFML to see what the robot sees.
• C++ Implementation: Optimized for low-latency obstacle detection.

Building & Running:

cd lidar
mkdir build && cd build
cmake ..
make
./ld19_lidar

A dedicated Python module for manual teleoperation.

• Gamepad Support: Designed for the Logitech F710 controller.
• Plug & Play: Automatically detects the controller and maps inputs to steering and throttle commands.
• Simple Interface: Easy-to-understand Python script using evdev.