A major constraint in planetary exploration is the limited data budget. Modern spacecraft and rovers can collect far more data than they can transmit back to Earth, creating a need for scientific autonomy. While AI cannot replace scientists, autonomous systems can help prioritize valuable observations, discard flawed data and identify findings that may lead to new discoveries. These challenges are driving the integration of advanced AI into space missions. For more than a decade, my research group has focused on automating scientific inquiry to enable autonomous decision-making systems. Our work spans applications in industry, healthcare, rehabilitation and, most notably, space exploration.

Key contributions include:

• Automated analysis of Martian surface images to infer weather conditions
• Automated classification of lunar regolith
• Autonomous and teleoperated lunar navigation
• Unsupervised terrain classification using LiDAR data
• Automated municipal waste collection
• Post-operative monitoring following total knee arthroplasty

In this talk, Professor McIsaac will highlight our successes, lessons learned and near-term research goals as we continue advancing AI-enabled autonomy for scientific discovery.