Hyperspectral Data Breakthrough & World-First Latency Achievement
CogniSAT-6 is the first mission to deliver near real-time insights from hyperspectral data over ISL (Inter-Satellite Links)—a challenge attempted since the early 2000s. The system has achieved an end-to-end latency of just 10 minutes, from image capture to insight delivery, with a goal of sub-5 minutes in the coming weeks—a world first in Earth Observation at this scale. This capability enables daily, real-time Earth insights at an unprecedented level.
AI-Driven Maritime Monitoring & Domain Awareness
The satellite’s onboard SPACE:AI has detected and analyzed hundreds of vessels across key maritime regions, providing real-time insights to improve domain awareness and security. Ubotica’s baseline maritime monitoring model is already state-of-the-art, and ongoing fine-tuning continues to enhance detection accuracy. The system is also capable of identifying dark ships—vessels operating without active transponders—by fusing onboard insights with AIS (Automatic Identification System) data.
AI-Powered Disaster Response & Environmental Monitoring
CogniSAT-6 processes and delivers real-time insights daily, demonstrating the scalability and reliability of AI-powered satellite autonomy. For example, NASA JPL’s Surface Water Insights for Flood Tracking (SWIFT) AI model estimated 21% flooding in an observed area during a Valencia flood event. Additionally, Ubotica collaborated with NASA JPL to deploy an AI-powered smoke plume detection system, identifying wildfire activity within minutes to enable faster emergency response.
AI Evolution: Expanding the Onboard Application Ecosystem
CogniSAT-6 has executed more than 20 onboard AI processing applications by NASA JPL, including thermal anomaly segmentation, surface water segmentation, ship detection, and more. Ubotica has delivered tens of software updates, continuously improving SPACE:AI, enhancing model performance, and adding new operational features. The fine-tuned version of the maritime monitoring model achieves even greater accuracy, demonstrating the continuous evolution of AI in orbit.
Next Steps: Advancing AI Autonomy & Dynamic Targeting
In the coming months, Ubotica will deploy and test new autonomy enhancements, including autonomous scheduling, Dynamic Targeting, and Tip-and-Cue—techniques that have had limited demonstrations but have never been integrated into a single system, let alone a CubeSat. These advancements will further push the boundaries of real-time, autonomous Earth Observation.
Looking Ahead
As CogniSAT-6 enters its second year in orbit, Ubotica remains committed to advancing satellite autonomy and real-time Earth observation. SPACE:AI enables satellites to Think. See. Act. autonomously, allowing them to analyze their environment, extract key insights, and take decisive actions in real-time. The achievements of the past year underscore the transformative potential of integrating AI into space-based platforms, paving the way for more responsive, intelligent, and autonomous satellite systems in the future.