A modular AI-driven framework integrating failure prediction, anomaly detection, and a decision support system for CubeSat platforms (1U–6U). Combines XGBoost, Time-Series Transformers, LSTM Autoencoders, and a PPO reinforcement learning agent with SHAP/LIME explainability throughout. Achieves 89.5–94.1% accuracy across platform classes. Aligned with UNOOSA and COSPAR guidelines.
An ongoing initiative developing a digital twin framework for precision agriculture, leveraging satellite remote sensing and AI-based environmental modelling to support real-time crop monitoring, resource optimization, and climate-resilient farming decisions.
Ongoing research into fixed-wing UAV autonomous systems, focusing on onboard AI decision-making, flight control, and mission planning under resource constraints — extending our core expertise in edge AI inference to aerial platforms.
10 papers accepted at the IAF Global Space Conference on Climate Change (GLOC 2026, Kigali). Presentation postponed; papers remain accepted and will be presented when rescheduled.
Each track operates independently with a dedicated lead, contributing to our shared research output. Tap any track to see its members.