Designed for drone research and real-world flight validation, the P600 research UAV development platform places greater emphasis on system-level integration. It combines the Prometheus and SpireCV software ecosystems and integrates flight control, visual recognition, LiDAR mapping, path planning, gimbal control, positioning and navigation, and ground-station debugging. This enables developers to complete the full workflow—from algorithm deployment to field testing—on a single platform.

On the hardware side, the P600 is equipped with a 40× zoom electro-optical gimbal and a 3D LiDAR. It supports long-range target detection and tracking, autonomous landing, 3D LiDAR SLAM mapping, and EGO-Swarm path planning and obstacle avoidance. With Prometheus Ground Station Professional, developers can view flight status in real time, execute missions, and tune key parameters, reducing the barrier to real-aircraft debugging and secondary development.

This latest P600 configuration further improves the field experience by upgrading key components including the G16 integrated video-and-data-link remote controller, dual-antenna high-precision RTK, and MID-360S 3D LiDAR. The platform is well suited to drone courses, research experiments, algorithm validation, and project prototyping.

01 Three Core Configuration Upgrades

When a research UAV moves into field testing, the communication link, positioning accuracy, and environmental perception directly affect the stability of algorithm tests and the quality of collected data. The P600 therefore introduces targeted upgrades to the remote controller, RTK, and LiDAR.

G16 Integrated Video-and-Data-Link Remote Controller

The new P600 is supplied with the G16 integrated video-and-data-link remote controller, bringing flight control, mission operations, video transmission, and data communication together in one terminal.

The combined video and data link reaches up to 30 km. The controller has a 7-inch sunlight-readable display and provides 9–10 hours of battery life. Operators can view flight status and video in real time while controlling the aircraft with the joysticks—making the flight visible and allowing what is seen to be controlled directly.

Integrating video transmission, telemetry, and remote-control links also simplifies equipment deployment and mission preparation, making the system suitable for classroom demonstrations, algorithm testing, and routine field flights.

Dual-Antenna High-Precision RTK

The P600 integrates an all-constellation, all-frequency positioning module and a LoRa data radio. It provides centimeter-level positioning without requiring RTCM transmission configuration. Its built-in battery eliminates the need for an external power supply, enabling flexible, ready-to-use deployment.

MID-360S 3D LiDAR

The MID-360S provides a 360° horizontal field of view and a maximum vertical field of view of 59°. At 10% target reflectivity, its maximum detection range reaches 40 m, while the near blind zone is only 0.1 m. It operates in challenging conditions including low light and strong light. With the MID-360S, the P600 gains an all-around, blind-spot-free field of view that supports seamless operation across different scenarios.

Combined with the onboard computing platform and related algorithms, the P600 can be used to develop 3D LiDAR SLAM mapping, obstacle perception, spatial positioning, and path-planning capabilities, providing perception data for autonomous-flight experiments outdoors and in complex spaces.

Compared with a vision-only solution, 3D LiDAR is less affected by changes in illumination, further extending the platform’s experimental capabilities in low-light environments, scenes with sharp brightness transitions, and areas with insufficient visual texture.

02 From Visual Perception to Autonomous Obstacle Avoidance

The P600 integrates an electro-optical gimbal, 3D LiDAR, RTK, and an onboard computing platform, supporting complete development workflows spanning target perception, spatial mapping, mission planning, and flight control.

Long-Range Target Detection and Tracking

YOLOv5, loaded with the general-purpose COCO dataset, is used for target detection. NanoTrack then locks onto and tracks the selected target. Without zoom, targets can be recognized and tracked at up to 30 m. The GX40’s 40× optical and digital zoom extends detection and tracking to much longer distances. During tracking, Prometheus Ground Station Professional can switch the dynamic target-following distance, with a default of 10 m.

QR-Code-Guided Autonomous Landing

Using SpireCV for accurate QR-code detection and position estimation, the GX40 can precisely guide the aircraft to land on a QR marker from different altitudes. The center of the electro-optical gimbal camera can be aligned with the QR-code center to within 30 cm.

3D LiDAR SLAM Mapping

The P600 uses the MID-360S to acquire environmental point clouds and completes localization and 3D map construction on its onboard computing platform.

Developers can view real-time point clouds, the aircraft position, and mapping results in Prometheus Ground Station Professional, making it easier to analyze LiDAR data, localization status, and algorithm performance.

This capability can be used for 3D mapping experiments on campuses, in industrial parks, forests, and other complex spaces, and it can provide the map foundation for subsequent path planning and autonomous obstacle avoidance.

Path Planning and Obstacle Avoidance

MID-360S publishes point-cloud and position data for the EGO-Swarm path-planning algorithm, significantly improving the aircraft’s autonomy in dynamic environments. The P600 can interpret its surroundings in real time, flexibly avoid moving obstacles, and quickly adapt to dynamically updated target points. With a maximum speed of 2 m/s, it demonstrates strong agility and adaptability in unknown and complex operating environments, improving mission efficiency and enabling diverse tasks to be completed accurately.

03 Software, Computing Power, and Ground Station Working Together

Allspark2 Onboard Computing Platform

The P600 carries the compact Allspark2 edge computer, built around an NVIDIA Jetson Orin NX module and delivering up to 100 TOPS of computing performance.

The onboard computer is compact and lightweight and comes with the relevant development environment preinstalled. It can run visual-recognition, point-cloud processing, SLAM mapping, and path-planning algorithms, reducing the time developers spend on environment setup and basic adaptation.

Prometheus V2 Software System

Prometheus V2 is built on ROS and the open-source Prometheus framework and provides access to UAV state, sensor data, and control interfaces.

Developers can read positioning data, flight mode, battery status, and IMU data, and can call position, velocity, acceleration, and attitude control interfaces to connect their own algorithms to the flight-control workflow.

The system also provides flight-safety checks. When certain abnormal states are detected, it can command an automatic landing to reduce the risk of a crash during real-aircraft development. This safety check is not an obstacle-avoidance function; it is intended primarily to handle abnormal flight states.

SpireCV Vision Library

SpireCV is an onboard real-time image-processing SDK for intelligent unmanned systems. It provides modules for video algorithms, video recording and streaming, camera and gimbal control, and target detection, recognition, and tracking.

Developers can quickly call existing vision functions through the available interfaces, or integrate their own detection, recognition, and tracking algorithms for secondary development of drone missions.

PrometheusGroundStation

PrometheusGroundStation is a Qt-based human-machine interface that communicates with the aircraft over TCP/UDP.

The ground station displays UAV status, positioning information, live flight video, and point-cloud maps in real time. It also provides one-click takeoff and landing, position control, and mission operations. Algorithm results, sensor status, and control commands can be viewed in the same interface, helping developers locate problems and tune parameters.

04 From Simulation Integration to Real-Flight Validation

Drone algorithm development typically passes through several stages: simulation validation, code deployment, system integration, and field testing. The P600 brings these stages into one software-and-hardware system, reducing repeated adaptation between different platforms.

Developers can first validate 3D LiDAR SLAM, path planning, and mission logic in the supplied simulation demos, and then deploy the algorithms to the Allspark onboard computing platform. After configuring interfaces and parameters, they can use Prometheus Ground Station Professional to view real-time data and carry out real-flight tests including hovering, mapping, obstacle avoidance, target tracking, and guided landing.

Using the same Prometheus interfaces and development framework from simulation through real flight also makes it easier to compare algorithm behavior in different environments and progressively complete debugging and optimization.

05 For Research, Teaching, and Prototype Development

The P600 is suitable for development and validation in the following areas:

· Drone-course teaching: Experiments in flight control, ROS development, visual recognition, LiDAR SLAM, and path planning.

· Research algorithm validation: Real-flight testing of target detection and tracking, QR-code-guided landing, 3D mapping, path planning, and dynamic obstacle avoidance.

· System-integration experiments: Coordinated development involving the flight controller, onboard computer, RTK, LiDAR, electro-optical gimbal, and ground station.

Project prototype development: Validation of mission workflows, algorithm performance, and software/hardware interfaces to provide an operational technical prototype for subsequent project development.