AEROSPACE MISSIONS
Space Exploration Missions enables anyone to use our simulator to develop autonomous AI algorithms for exploration devices deployed on extraterrestrial planets. These algorithms can perform tasks such as perception, navigation, and obstacle avoidance.
REALSIM
MULTIPLE TASKs
EXPLORATION
Surface Exploration Tasks involve autonomous rovers conducting long-term traversal, analysis, and decision-making across planetary landscapes. These missions demand robust planning and perception systems capable of navigating rough, dusty, and visually ambiguous terrains. Key difficulties include generating terrain-aware routes under uncertainty, handling degraded sensor performance due to environmental extremes (e.g., low light or dust storms), and maintaining mission continuity during communication blackouts, where real-time human intervention is impossible.
LANDING
Landing Tasks deal with the precise delivery of robotic systems or payloads from orbit to planetary surfaces. A prime example is the Mars Perseverance rover, which utilized real-time vision-based hazard detection during descent to land safely in Jezero Crater. Challenges here include simulating atmospheric entry dynamics, accurately modeling parachute deployment and retro-thruster control, and ensuring autonomous hazard avoidance for pinpoint landings on unpredictable terrain—all within seconds and with minimal ground intervention.
ORBITAL
Orbital Tasks focus on the operation and simulation of satellites orbiting planetary bodies, enabling wide-area observation and support for downstream missions. These tasks involve training autonomous orbiters to navigate complex gravitational fields and coordinate multi-sensor data acquisition. Major challenges include achieving high-precision orbital control, transforming raw orbital imagery or radar data into terrain intelligence, and synchronizing observation schedules with surface or landing activities—despite limited communication windows and onboard processing power.
FEATURES
SIMULATION
SURVEILLANCE
MAINTENANCE
WORK FLOW
Users can use our tools to build their own commercial space projects. In addition to the classic space project cases we provide, you can still continue to build your own commercial space projects.
Operational Design Domain
Define the environmental requirements for the robot to work properly.
Pre-trained Model
Choose pre-trained models for perception, planning, and control models.
Fine-Training and Optimizations
Fine-tune model parameters and optimize algorithm engineering.
Defining Onboard Sensors
Choose from a variety of sensors.
Kinetic Calibration
Define motor parameters and calibrate dynamic model.
Simulation and Data Collection
Validate algorithms in simulation and collect your own training data.
Simulation Data
Sensor data, trajectory data, inference result data
Simulation and Metric Test
Perform metric evaluation on the standard test set to observe the core indicators of the algorithm.
REALTIME MARS
For information about our launch services,contact sales.aero@starcloneinc.com