Informative Path Planning Toward Autonomous Real-World Applications

Gathering information from the physical world is critical for applications such as scientific research, environmental monitoring, search and rescue, defense, and disaster response. Autonomous robots provide significant advantages for information gathering, particularly in situations where human access is constrained, hazardous, or impractical. By leveraging intelligent algorithms, these robots can efficiently collect data, enhancing decision-making and accelerating insights. Informative Path Planning (IPP) plays a key role in maximizing the effectiveness of robotic information gathering by generating paths that optimize data collection while respecting operational constraints. This thesis advances autonomous information gathering by addressing three key challenges: (1) solving IPP problems in high-dimensional spaces with complex sensor constraints, (2) incorporating real-world disturbances and risk into the planning framework, and (3) improving and leveraging world belief models. First, we in troduce IA-TIGRIS, an adaptive, incremental sampling-based planner that efficiently computes long-horizon information-gathering paths while respecting vehicle motion constraints and non-trivial sensor footprints. We validate the planner through ex tensive simulations and real-world field tests on multiple unmanned aerial vehicle (UAV) platforms. Second, we develop a real-time wind field estimation method using onboard UAV measurements, a time-optimal path planner for UAVs in wind, and a deep learning-based energy risk assessment framework to quantify flight risk under uncertain environmental conditions. Third, we propose a new belief representation for search and tracking, along with planning approaches that incorporate human-inspired heuristics and predictive belief models. The proposed methods are validated through a combination of simulation studies and extensive field deployments, contributing to the development of robust, real world-ready autonomous systems. We demonstrate their effectiveness across diverse planning scenarios and on multiple UAV platforms, including both fixed-wing and multirotor systems. While the primary focus is on robotic information gathering, the underlying algorithms generalize to broader applications such as robotic exploration, active perception, target tracking, and multi-agent coordination.