ROBUST PLANNING FOR DYNAMIC TENSEGRITY STRUCTURES THE PROPOSED PROJECT WILL RESEARCH AND DEVELOP HIGH-PERFORMANCE MOTION PLANNERS FOR DYNAMIC TENSEGRITY STRUCTURES WHICH CAN BE USED AS ROBOTIC MOBILITY AND MANIPULATION PLATFORMS. THE FOCUS WILL BE ON EFFECTIVE PLANNING UNDER NON-TRIVIAL DYNAMICS COMPLEX CONTACTS WITH A PLANETARY SURFACE AS WELL AS UNCERTAINTY IN LOCATION ESTIMATION AND ACTUATION. IN THE ABOVE CONTEXT THIS WORK WILL AIM TO PROVIDE METHODS WITH STRONG PERFORMANCE GUARANTEES REGARDING THE QUALITY AND ROBUSTNESS OF TRAJECTORIES FOR TENSEGRITY STRUCTURES. THIS EFFORT WILL PROGRESS THE STATE-OF-THE-ART IN PLANNING FOR TENSEGRITIES BY PROVIDING PLANNERS THAT REASON ABOUT COMPLEX CONTACTS FOR SOFT ROBOTICS PLATFORMS. THESE TECHNIQUES SHOULD ALSO BE PRACTICAL AND COMPUTATIONALLY EFFICIENT AND THEY WILL BE EVALUATED ON REAL TENSEGRITY PLATFORMS CONSIDERED AS POSSIBLE PLANETARY EXPLORATION PROBES. CLASSICAL LOCOMOTION AND MANIPULATION SYSTEMS EMPLOY LEGS HANDS OR WHEELS BUT HAVE LIMITED DEFORMATION. TENSEGRITY IS A GOOD CANDIDATE TO CONSTRUCT ROBOTS WITH ARBITRARY FORMS THAT ADAPT EFFICIENTLY TO HIGHLY UNSTRUCTURED AND CLUTTERED ENVIRONMENTS. TENSEGRITY ROBOTS ARE LIGHTWEIGHT IMPACT TOLERANT AND OFFER UNIQUE MODES OF LOCOMOTION AS THEY CAN ROLL CRAWL OR GALLOP DEPENDING ON CONSTRUCTION AND NEED. THEY ARE ESPECIALLY WELL SUITED FOR TASKS THAT REQUIRE LOW-WEIGHT ENERGY EFFICIENT PLATFORMS AS THEY ARE TYPICALLY MADE OF TUBES/RODS AND CABLES/ELASTIC LINES. THE USE OF ELASTIC TENSILE COMPONENTS AND DYNAMICAL GAITS ENABLES EFFICIENT MOVEMENT. THE ABOVE PROPERTIES MAKE TENSEGRITY STRUCTURES APPROPRIATE FOR SPACE APPLICATION AS THEY INTRODUCE SMALL WEIGHT OVERHEAD RELATIVE TO THE SCIENCE PAYLOAD UPON ATMOSPHERIC ENTRY ACT AS AIR-BAGS UPON LANDING PROTECTING THE KEY EQUIPMENT AND ARE ABLE TO PROVIDE SURFACE MOBILITY AND MANIPULATION CAPABILITIES. THIS IMPLIES THAT FUTURE MISSIONS CAN BE CHEAPER AND CAN UTILIZE NEW WAYS OF INTERACTING WITH PLANETARY SURFACES. DESPITE THEIR BENEFITS THE CONTROL OF TENSEGRITY ROBOTS IS CHALLENGING DUE TO THE INFINITE DIMENSIONALITY AND DEPENDENCE ON NON-LINEAR DYNAMICS. IN PARTICULAR TWO KEY ASPECTS OF TENSEGRITY STRUCTURES MAKE THEM HARD TO OPERATE USING CLASSICAL CONTROL METHODOLOGIES: (I) A FORCE APPLIED ON A TENSEGRITY STRUCTURE CAUSES NON-LINEAR RESPONSES AND (II) OSCILLATORY BEHAVIORS ARISE FROM INTERACTIONS WITH THE ENVIRONMENT. THESE ISSUES HAVE LIMITED CURRENT PATH PLANNING TO SOLUTIONS TO THE GENERATION OF QUASI-STATIC PATHS. EXISTING SOLUTIONS EITHER DEAL WITH GATE GENERATION AND DO NOT ADDRESS GLOBAL PATH PLANNING OR ONLY PROVIDE QUASI-STATIC COLLISION-FREE PATHS. THIS PROPOSAL WILL CREATE NEW TECHNOLOGIES THAT OVERCOME THESE CHALLENGES SO THAT FUTURE NASA MISSIONS CAN TAKE ADVANTAGE OF THEM. IN PARTICULAR THE CONSIDERED SCENARIOS INVOLVE TENSEGRITY PROBES AFTER LANDING WHICH ARE ABLE TO TRAVERSE ROUGH TERRAIN DESCEND EXTREME SLOPES AND ALLOW SCIENTIFIC EQUIPMENT TO INTERACT AND MANIPULATE OBJECTS IN THE ENVIRONMENT SUCH AS COLLECTING GROUND SAMPLES OR GRASPING ROCKS. THE PROJECT WILL PROVIDE AN IMPORTANT APPLICATION DOMAIN FOR RECENT RESULTS ACHIEVED BY THE PI'S GROUP REGARDING PRACTICAL METHODS THAT CAN ACHIEVE ANYTIME PERFORMANCE WHEN PLANNING FOR SYSTEMS WITH NON-TRIVIAL DYNAMICS AND ESPECIALLY UNDER THE PRESENCE OF SIGNIFICANT NON-GAUSSIAN UNCERTAINTY. THE FOCUS WILL BE ON EFFICIENT BELIEF-SPACE PLANNING FOR TENSEGRITY ROBOTS WHICH REASONS ABOUT THE EFFECTS OF CONTACTS AND COMPUTES ROBUST TRAJECTORIES THAT MAXIMIZE THE PROBABILITY OF REACHING A DESIRED CONFIGURATION. THIS WORK HAS THE PROSPECT OF IMPACTING NASA'S FUTURE MISSION AS TENSEGRITY STRUCTURES ARE PROMISING SPACE EXPLORATION PROBES. THE METHODS WILL BE EVALUATED INITIALLY USING PHYSICS-BASED SIMULATION BY INTEGRATING OPEN-SOURCE SOFTWARE BY THE INTELLIGENT ROBOTICS GROUP AT NASA AMES. THROUGH INTERACTIONS WITH THE GROUP IT WILL ALSO BE POSSIBLE TO TEST THE METHODS ON THE PHYSICAL PLATFORMS UNDE