IN THIS PROJECT WE PROPOSE TO EXPLORE THE APPLICATION OF MARKOV BRAINS (NATURAL COGNITIVE ALGORITHMS) TO PERFORM ON-BOARD IMAGE PROCESSING AND SHAPE MODEL GENERATION. MARKOV BRAINS (EDLUND ET AL. 2011) REPRESENT PROBABILISTIC OR DETERMINISTIC NETWORKS OF DIGITAL NEURONS THAT CAN MAKE DECISIONS BASED ON A COMBINATION OF SENSORY DATA AND INTERNAL STORED REPRESENTATIONS. THE COGNITIVE ALGORITHM WE PROPOSE MERGES TWO TECHNOLOGIES THAT HAVE ALREADY BEEN VALIDATED BY THEMSELVES IN DIFFERENT CONTEXTS BUT HAVE NEVER BEEN PUT TOGETHER FOR AUTONOMOUS NAVIGATION IN DATA-LIMITED ENVIRONMENTS: - THE EVOLUTION/TRAINING OF AN ARTIFICIAL BRAIN FOR OBJECT CLASSIFICATION BASED ON A SEQUENCE OF PATCH MEASUREMENTS AND - THE GENERATION OF 3D SHAPES FROM INCOMPLETE DATA USING STORED REPRESENTATIONS. THESE TECHNIQUES ARE ESSENTIAL TO DEVELOPING NEW TECHNOLOGIES AND CAPABILITIES TO ENABLE DEEP SPACE MISSIONS WITH SMALL SPACECRAFT WHILE MAINTAINING THE AFFORDABILITY OF SMALL SPACECRAFT SYSTEMS (TECHNOLOGY TOPIC AREA 3).