Python SDK Overview¶

RflySim's Python control interface is uniformly integrated into RflySimSDK, located under [Installation Directory]\RflySimAPIs\RflySimSDK\. This SDK provides comprehensive capabilities for drone simulation control, visual sensing, swarm collaboration, and health management development.

Module Architecture¶

RflySimSDK Module Structure

Core Module Quick Reference¶

ctrl — Flight Control ¶

MAVLink communication core with PX4/APM flight controllers, supporting single-vehicle and multi-vehicle control.

Interface File	Core Class	Function
PX4MavCtrlV4	`PX4MavCtrler`	MAVLink control core: arm/takeoff/land/Offboard/waypoint
PX4MavCtrlV4ROS	`PX4MavCtrlerROS`	ROS-integrated PX4 controller
DllSimCtrlAPI	`DllSimModel`	CopterSim DLL integrated model interaction
DllSimCtrlAPIROS	`DllSimModelROS`	DLL model ROS integration
api	`RflyCtrl`	High-level simplified API
ReqCopterSim	`ReqCopterSim`	CopterSim simulation initialization request
EarthModel	`EarthModel`	WGS84/UTM/NED coordinate transformation
IpManager	`IpManager`	IP address allocation and management
RflyRosStart	`RflyRosStart`	Automatic ROS environment startup

comm — Network Communication ¶

Cluster network communication simulation, supporting UDP broadcast/unicast and Redis message middleware.

Interface File	Core Class	Function
NetSimAPIV4	`NetSimAPI`	Custom cluster networking simulation
NetUavAPI	`NetUavAPI`	Multi-UAV cluster networking communication
RedisUtils	`RedisUtils`	Redis publish-subscribe utility

vision — Vision Sensing ¶

Vision sensor data acquisition and control, supporting RGB images, depth maps, point clouds, and LiDAR.

Interface File	Core Class	Function
VisionCaptureApi	`VisionCaptureApi`	Image/depth/point cloud data acquisition
CameraCtrlApi	`CameraCtrlApi`	Camera parameter control (resolution/FOV/pose)
ScreenCapApiV4	`ScreenCapApiV4`	Screen capture acquisition
Open3DShow	`Open3DShow`	Real-time Open3D point cloud visualization
RflyRosCtrlApi	`RflyRosCtrlApi`	ROS vision data publishing and control

ue — Engine Control ¶

Interaction and control with RflySim3D/RflySimUE5 3D engines.

Interface File	Core Class	Function
UE4CtrlAPI	`UE4CtrlAPI`	Engine console commands, scene interaction
UEMapServe	`UEMapServe`	Map loading and scene switching service
RobotDogStart	`RobotDogStart`	Robot dog simulation startup

swarm — Swarm Control ¶

Large-scale multi-vehicle swarm management, distributed simulation, and heterogeneous platform adaptation.

Interface File	Core Class	Function
VehicleApi	`VehicleApi`	Single-vehicle state subscription and control encapsulation
distSimCtrlAPI	`DistSimCtrlAPI`	Distributed simulation node control
Crazyflie	`Crazyflie`	Crazyflie micro-drone adaptation
CrazySwarm	`CrazySwarm`	CrazySwarm swarm framework
MavRflyShell	`MavRflyShell`	MAVLink command-line tool
RflyADBLib	`RflyADBLib`	ADB remote debugging library
RJ45_px6x	`RJ45_px6x`	Ethernet-based flight controller communication (Pixhawk 6x)

phm — Health Assessment ¶

Automated testing, fault injection, flight data logging, and safety evaluation.

Interface File	Core Class	Function
AutoMavCtrl	`AutoMavCtrl`	Automated test process management
AutoMavDB	`AutoMavDB`	Test case database management
AutoMavCmd	`AutoMavCmd`	Standardized control sequence parsing
Ass	`Ass`	Safety and reliability assessment
DeBug	`DeBug`	Debugging assistance tool
QGCCtrlAPI	`QGCCtrlAPI`	QGC ground station control interface
AutoREG	`AutoREG`	Automatic registration and discovery
AutoVisConf	`AutoVisConf`	Visual configuration automation
MavDataRec	`MavDataRec`	MAVLink flight data recording

Quick Start¶

Minimal Example: Single-Vehicle Takeoff and Landing¶

from RflySimSDK.ctrl.PX4MavCtrlV4 import PX4MavCtrler
import time

# Create control instance (connect to vehicle ID=1)
mav = PX4MavCtrler(ID=1)

# Send global start signal
mav.sendStartMsg(copterID=-1)
mav.waitForStartMsg()

# Arm → Takeoff → Hover → Land
mav.SendMavArm(True)
mav.SendMavTakeoff(5)   # Take off to 5m
time.sleep(10)
mav.SendMavLand()

Multi-Vehicle Cluster Example¶

from RflySimSDK.ctrl.PX4MavCtrlV4 import PX4MavCtrler

# Create control instances for 3 drones
mavs = [PX4MavCtrler(ID=i+1) for i in range(3)]

# Global startup
mavs[0].sendStartMsg(copterID=-1)
for m in mavs:
    m.waitForStartMsg()

# Individual takeoff
for m in mavs:
    m.SendMavArm(True)
    m.SendMavTakeoff(5)

Vision Sensor Image Capture¶

from RflySimSDK.vision.VisionCaptureApi import VisionCaptureApi
import cv2

# Initialize visual capture
vis = VisionCaptureApi()
vis.jsonLoad("Config.json")

# Retrieve image
img = vis.getImg(0)
cv2.imshow("Camera", img)
cv2.waitKey(0)

Environment Configuration¶

The SDK is pre-configured in RflySim's built-in Python environment and can be directly imported.

To use in a custom Python environment:

# Method 1: Run the registration script
python RflySimSDK/ReLabPath.py

# Method 2: Manually set environment variables
set PYTHONPATH=%PSP_PATH%\RflySimAPIs;%PYTHONPATH%

# Method 3: Install dependencies via pip
pip install pymavlink numpy opencv-python

Feature Highlights¶

Feature	Description
Full-stack Control	Complete flight control chain from low-level MAVLink to high-level API
Multiple Communication Modes	Four connection methods: UDP / Serial / Direct / Redis
Real-time Simulation	Supports HITL hardware-in-the-loop and SITL software-in-the-loop simulation
Swarm Collaboration	Supports large-scale swarm control of hundreds of drones
Visual Perception	RGB images, depth maps, LiDAR point clouds, IMU data acquisition
Fault Injection	Motor/sensor fault simulation and automated batch testing
ROS Integration	Native support for ROS1/ROS2 Topic publishing and subscription
Cross-platform	Compatible with Windows, Linux (WSL/Docker) environments