Carla Based Urban Drone Simulation

Simulate urban drone flights in CARLA Simulator with ease—ideal for computer vision, urban analytics, and autonomous systems projects.

Perfect for vision-based projects in urban environments:

• Test and develop algorithms for object detection, tracking, and scene understanding
• Simulate smart city scenarios with realistic traffic, weather, and lighting
• Collect synthetic data for deep learning and AI models
• Experiment with drone navigation and perception in complex cityscapes

Easily configure drone camera, speed, number of cars, weather, and time of day. Visualize the drone's camera in real time with OpenCV.

---

Features

• Modular Python class for drone simulation in CARLA
• Real-time RGB camera view with OpenCV
• User-configurable:
  • Drone speed, height, and camera angle
  • Number of moving and parked cars
  • Video quality (360p, 480p, 720p, 1080p)
  • Weather: clear or rainy
  • Time of day: day or night
  • Traffic light timing
  • FPS overlay (optional)
• Real-time access to drone speed and road ID
• Simulated battery drain and speed reduction as battery drops
• Console output of battery and speed in real time
• Clean, extensible codebase

---

Installation

1. Install Python (3.8+ recommended)

Download from python.org.

2. Install CARLA Simulator

• Download CARLA from carla.org/downloads
• Extract and run the simulator:
  • On Windows: double-click CarlaUE4.exe or run from terminal:

    CarlaUE4.exe -carla-port=YOUR_DESIRED_PORT

  • On Linux:

    ./CarlaUE4.sh -carla-port=YOUR_DESIRED_PORT

• Wait for the CARLA window to fully load before running the Python script.

3. Install Python dependencies

Create a requirements.txt file with:

carla
opencv-python
numpy
pandas

Install with pip:

pip install -r requirements.txt

Note:

• The carla Python package must match your CARLA server version. If you have issues, see CARLA Python API installation guide.

---

Usage

1. Edit the configuration in DroneSimulation.py (bottom of the file):

config = DroneSimConfig(
    drone_view_angle=-55.0,      # Camera pitch angle
    drone_height=60.0,           # Drone flight height
    drone_speed=1.0,             # Drone speed (m/s)
    num_moving_cars=10,          # Number of moving cars
    num_parked_cars=5,           # Number of parked cars
    traffic_lights={"all": (5.0, 12.0, 2.0)},
    carla_host="localhost",
    carla_port=4000,
    video_quality="good",       # "low", "medium", "good", "very good"
    display_fps=True,            # Show FPS overlay
    time_of_day="night",        # "day" or "night"
    weather_mode="rainy"        # "clear" or "rainy"
)
sim = DroneSimulation(config)
sim.run()

2. Run the simulation for test:

python test.py

3. Control:

• Press q in the OpenCV window or Ctrl+C in the terminal to stop.
• The console will show the drone's current position, battery percentage, and speed in real time.
• When the battery drops below 30%, the drone will slow down automatically.

---

Example Console Output

Drone position: x=10.00, y=20.15, z=60.00, yaw=-55.0, cam_yaw=0.0 | Battery: 98.50% | Speed: 1.00 m/s
Drone position: x=..., ... | Battery: 28.00% | Speed: 0.93 m/s

---

Configuration Options

Parameter	Type	Description	Example
drone_view_angle	float	Camera pitch angle (degrees)	-55.0
drone_height	float	Drone flight height (z)	60.0
drone_speed	float	Drone speed (m/s)	1.0
num_moving_cars	int	Number of moving cars	10
num_parked_cars	int	Number of parked cars	5
traffic_lights	dict	Traffic light timing (all or by ID)	{"all": (5,12,2)}
carla_host	str	CARLA server host	"localhost"
carla_port	int	CARLA server port	4000
video_quality	str	Camera resolution: "low", "medium", "good", "very good"	"good"
display_fps	bool	Show FPS overlay in OpenCV window	True
time_of_day	str	"day" or "night"	"night"
weather_mode	str	"clear" or "rainy"	"rainy"
battery/speed	auto	Simulated battery drains over time, speed drops as battery gets low	(automatic)

---

Example requirements.txt

carla
opencv-python
numpy
pandas

---

Troubleshooting

• CARLA connection errors:
  • Make sure the CARLA server is running and fully loaded before starting the Python script.
  • Ensure the carla_port matches the port used to launch CARLA.
  • The Python carla package version must match the CARLA server version.
• No OpenCV window:
  • Make sure you have a display (not running headless).
• Performance:
  • Lower the video quality or number of cars for better performance.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.