**AWARDS ISSUED PRIOR TO JANUARY 20, 2025, WERE FUNDED UNDER PREVIOUS ADMINISTRATIONS AND MAY NOT REFLECT THE PRIORITIES AND POLICIES OF THE CURRENT ADMINISTRATION.** DECADES OF AGRICULTURAL RESEARCH SHOW THAT DEPLOYING SENSOR NETWORKS TO MONITOR SOIL MOISTURE, LEAFWETNESS, TEMPERATURE, OR OTHER VARIABLES OF INTEREST ON A FARM REDUCES WASTE AND MAINTAINS ORIMPROVES YIELDS. DESPITE THE ADVANTAGES, MANY FARMS DO NOT UTILIZE SENSOR DATA IN THEIR LANDMANAGEMENT DECISIONS BECAUSE DEPLOYING EXTENSIVE COMMUNICATIONS AND POWER INFRASTRUCTURE TOCOLLECT SENSOR DATA INCURS PROHIBITIVE COST AND MAINTENANCE REQUIREMENTS.UNMANNED AERIAL AND GROUND VEHICLES (UAVS AND UGVS, RESPECTIVELY) HAVE DEMONSTRATED VERSATILITYACROSS MANY ROUTINE AGRICULTURAL AND EXPLORATION TASKS, BUT AUTONOMOUS OPERATION HAS NOT YET GAINEDTRACTION. ONE LIMITING FACTOR IS EQUIPMENT ENDURANCE REQUIREMENTS. FOR EXAMPLE, A TYPICAL QUADCOPTERUAV FLIES FOR APPROXIMATELY 30-60 MINUTES, AND UGVS RUN FOR 2-4 HOURS. THESE CAPABILITIES LEAD TOAN EFFECTIVE MOBILITY RANGE OF 2-3 KILOMETERS, BUT MOST US FARMS WOULD HAVE A SENSOR READING PLANTHAT WOULD REQUIRE A MOBILITY RANGE AN ORDER OF MAGNITUDE GREATER THAN THIS. TO SCALE IN-SITU SENSING TOREAL FARMS, WE NEED ADAPTIVE, RESOURCE-AWARE MISSION PLANNING TO MAKE THE MOST OF AVAILABLE FIELDROBOTICS TECHNOLOGIES AND OFFER DATA-DRIVEN INSIGHT TO LAND MANAGEMENT DECISIONS.WE PROPOSE TO DEVELOP NOVEL TECHNIQUES TO ACQUIRE MEASUREMENTS IN THE AREAS OF GREATEST INFORMATIONGAIN, THUS INCREASING THE EFFECTIVE SIZE OF THE FARM THAT CAN BE MONITORED WITH A GIVEN ROBOT TEAM. THISWILL ENABLE FARMERS TO PERSISTENTLY MONITOR THE HEALTH OF A FARM WHILE MINIMIZING THEIR COSTS. THISPROPOSAL WILL USE SIMULATION TECHNIQUES TO PROTOTYPE EFFECTIVE SENSOR SAMPLING POLICIES, FOLLOWED BYAN EVALUATION OF THE REAL-WORLD PERFORMANCE OF OUR ALGORITHMS.