Fast Frontier-based Information-driven Autonomous Exploration with an MAV

Abstract: Exploration and collision-free navigation through an unknown environment is a fundamental task for autonomous robots. In this paper, a novel exploration strategy for Micro Aerial Vehicles (MAVs) is presented. The goal of the exploration strategy is the reduction of map entropy regarding occupancy probabilities, which is reflected in a utility function to be maximised. We achieve fast and efficient exploration performance with tight integration between our octree-based occupancy mapping approach, frontier extraction, and motion planning-as a hybrid between frontier-based and sampling-based exploration methods. The computationally expensive frontier clustering employed in classic frontier-based exploration is avoided by exploiting the implicit grouping of frontier voxels in the underlying octree map representation. Candidate next-views are sampled from the map frontiers and are evaluated using a utility function combining map entropy and travel time, where the former is computed efficiently using sparse raycasting. These optimisations along with the targeted exploration of frontier-based methods result in a fast and computationally efficient exploration planner. The proposed method is evaluated using both simulated and real-world experiments, demonstrating clear advantages over state-of-the-art approaches.

• The paper presents a novel exploration strategy that combines frontier- and sampling-based methods to reduce map entropy.
• It leverages octree-based occupancy mapping and sparse raycasting to cut computational time to one-quarter of traditional methods.
• Experimental results in simulations and real-world environments validate enhanced exploration performance and efficient MAV navigation.

Fast Frontier-based Information-driven Autonomous Exploration with an MAV

This paper articulates a novel exploration strategy focusing on Micro Aerial Vehicles (MAVs), designed to achieve fast and efficient exploration in unknown environments. The primary objective is to minimize map entropy related to occupancy probabilities through a utility function. The strategy emerges from a hybrid approach combining frontier-based and sampling-based exploration methodologies, attaining improved computational efficiency and enhanced exploration performance.

The crux of the exploration strategy lies in the integration among octree-based occupancy mapping, frontier extraction, and motion planning, which circumvents the typical computational bloat seen in traditional methods. By leveraging implicit voxel clustering via the octree structure, the paper sidesteps the conventional requirement for frontier voxel clustering. Through continuous sparse raycasting, the exploration method evaluates candidate next-views based on a utility function that considers both map entropy and travel time. This methodological enhancement delivers a more computationally efficient planner.

The frontier-based method conducts operations in one-fourth of the traditional computational time when compared to state-of-the-art methods, demonstrating substantial advantages in both simulated and real-world environments. It notably minimizes path planning complexity by utilizing supereight, a volumetric map with effective octree structures for fast processing of occupancy probabilities and frontier detection. MAV deployments, both in virtual and physical settings, underscore this strategy's validity and efficiency.

Key to the proposed approach is the calculation and maximization of a utility function that emphasizes the trade-off between information gain (measured through map entropy) and travel time, resulting in a more purpose-driven navigation. The approach to evaluating and selecting candidate poses from frontier points ensures targeted and optimal explorative paths. In experimental simulations, testing in apartment, maze, and powerplant environments indicated superior performance in explored volume and time efficiency compared to pre-existing methods. Real-world testing further affirmed its feasibility and operational efficiency.

The paper contributes significantly to the theoretical underpinnings and practical applications of MAV navigation in areas such as mining, agriculture, and disaster response. It provides a robust exploration pipeline that supports on-board operations in MAVs, making it highly relevant for field applications. Looking forward, integrating multi-resolution mapping and dynamic trajectory planning stands as promising extensions for expanding the system's functionality and applicability. Additionally, the approach holds potential for adaptation to other sensor types, paving the way for broader application in diverse environments, including dynamic settings.