History-Aware Autonomous Exploration in Confined Environments Using MAVs

Abstract

Many scenarios require a robot to be able to explore its 3D environment online without human supervision. This is especially relevant for inspection tasks and search and rescue missions. To solve this high-dimensional path planning problem, sampling-based exploration algorithms have proven successful. However, these do not necessarily scale well to larger environments or spaces with narrow openings. This paper presents a 3D exploration planner based on the principles of Next-Best Views (NBVs). In this approach, a Micro-Aerial Vehicle (MAV) equipped with a limited field-of-view depth sensor randomly samples its configuration space to find promising future viewpoints. In order to obtain high sampling efficiency, our planner maintains and uses a history of visited places, and locally optimizes the robot's orientation with respect to unobserved space. We evaluate our method in several simulated scenarios, and compare it against a state-of-the-art exploration algorithm. The experiments show substantial improvements in exploration time (2× faster), computation time, and path length, and advantages in handling difficult situations such as escaping dead-ends (up to 20× faster). Finally, we validate the on-line capability of our algorithm on a computational constrained real world MAV.• Boosting the RRT planning performance by using a history of exploration potential as seeds. • Increasing the sampling efficiency by maximizing the

An Efficient Sampling-Based Method for Online Informative Path Planning in Unknown Environments

“…In a receding horizon (RH) fashion, a tree of views is sampled and the first segment of the best branch is executed. To escape local minima more efficiently, Witting et al [7] keep track of the planner's history as potential areas for reseeding the tree. Similarly, Selin et al [16] use the RH-NBV planner for local exploration and a frontier method for global goal selection.…”

“…Since the operational time of mobile robots is typically limited by the battery life, efficient computation of informative trajectories is of major importance. Traditional sampling-based online IPP approaches, based on repeatedly expanding a rapidly-exploring random tree (RRT) [1], [3], [7], [8], [9], iteratively sample feasible paths from the current robot pose, storing them in a tree structure, and execute the beginning of the best branch. However, these approaches face two important disadvantages.…”

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

“…RELATED WORK A considerable amount of work is available on autonomous exploration. The utility metrics that drive the exploration can be grouped into two categories, map entropy metrics [4] and unknown volume metrics [5]. These metrics can be utilised by either of the two main exploration strategies which are sampling-based and frontier-based exploration.…”

“…The NBV is found by growing an RRT to sample a position and yaw angle in free space. This method is improved in [5] where a history of visited locations is maintained to avoid being stuck at local minima.…”

3D Exploration and Navigation with Optimal-RRT Planners for Ground Robots in Indoor Incidents

“…Recently, several advances in exploration for aerial vehicles have been made, some examples are found in References [15][16][17]. However, ground robots are still necessary because of their larger autonomy and their higher payload capacity.…”