In order to cope with development needs in complex scenarios,PrometheusSim The simulation platform has been newly upgraded with six new core functions, covering full-link simulation capabilities from stand-alone to cluster, from perception to decision-making, allowing developers to verify cutting-edge algorithms anytime and anywhere.

Cluster formation

Smart formation | Multiple formations, switch at will

Dynamic formation switching：

It supports "one-line" and "triangular formation" preset formations, and realizes real-time formation transformation through the pilot-following method, with the error accuracy controlled within 0.1 meters.

Virtual structure algorithm:

Based on the group dynamics model, the cluster maintains formation integrity.

Full-dimensional obstacle avoidance

LiDAR and depth visual perception

LiDAR obstacle avoidance

360° scanning without blind spots:
Based on Livox Mid-360LiDAR, the simulation point cloud density reaches 200,000 points/second, supporting real-time modeling of dynamic obstacles.EGO-Plannerpath planning: Integrate the AMOVLABopen-source algorithm to achieve millisecond-level obstacle avoidance trajectory re-planning.

visual obstacle avoidance

Binocular stereo vision: Simulates the noise and lighting effects of real depth cameras, and supports ROS-driven obstacle convex hull detection.
multi-UAV collaborative obstacle avoidance： Supports simultaneous sensing of multiple UAVs to avoid cross-collision during group flight.
EGO-Plannerpath planning： Integrating the AMOVLABopen-source algorithm enables autonomous flight in complex environments.

2D LiDAR obstacle avoidance

Radar simulation adaptation: Simulate the raster mapping of devices such as Silan S3 in low-altitude environments, and plan obstacle avoidance and other functions.
EGO-Plannerpath planning： Integrating the AMOVLABopen-source algorithm enables autonomous flight in complex environments.
multi-UAV collaborative obstacle avoidance： Supports simultaneous sensing of multiple UAVs to avoid cross-collision during group flight.

Smart tracking

YOLOv5 dynamic target tracking

Deep learning optimization: The model is pre-trained in the UE4 virtual scene, and the recognition accuracy is increased to 98\%, supporting the tracking of fast-moving targets such as running and cycling.
Target recognition scheme adaptation: Click the target through the Prometheus ground control station or Image Window interface to start autonomous tracking, which is suitable for film and television tracking, emergency rescue and other scenarios.

Developer experience

The ultimate integration of Windows+WSL2

One-click deployment script

half an hour Complete the entire process from environment configuration to first flight.

Deep integration of Prometheus tool chain

Supports Ego-Planner, FAST-LIO and other algorithms for plug-and-play and compatibility ROS1/ROS2 Dual framework, seamless connection with real device development.

Resource Express

Wiki link:https://s.c1ns.cn/U0PTM