Next-Best-View planning for surface reconstruction of large-scale 3D environments with multiple UAVs

Abstract

In this paper, we propose a novel cluster-based Next-Best-View path planning algorithm to simultaneously explore and inspect large-scale unknown environments with multiple Unmanned Aerial Vehicles (UAVs). In the majority of existing informative path-planning methods, a volumetric criterion is used for the exploration of unknown areas, and the presence of surfaces is only taken into account indirectly. Unfortunately, this approach may lead to inaccurate 3D models, with no guarantee of global surface coverage. To perform accurate 3D reconstructions and minimize runtime, we extend our previous online planner based on TSDF (Truncated Signed Distance Function) mapping, to a fleet of UAVs. Sensor configurations to be visited are directly extracted from the map and assigned greedily to the aerial vehicles, in order to maximize the global utility at the fleet level. The performances of the proposed TSGA (TSP-Greedy Allocation) planner and of a nearest neighbor planner have been compared via realistic numerical experiments in two challenging environments (a power plant and the Statue of Liberty) with up to five quadrotor UAVs equipped with stereo cameras.

Multi-Robot Hybrid Coverage Path Planning for 3D Reconstruction of Large Structures

“…Some applications require achieving complete coverage using the various CPP techniques such as agricultural surveying, structure painting, lawn mowing, surveillance, geospatial mapping, object reconstruction, and floor cleaning. Generally in CPP, either the model is reconstructed in real time utilizing the robot's sensing capabilities (non-model based) [23], [27], or a reference model is provided in advance for the structure or the environment of interest (model-based) [10], [28]. Extensive reviews of the various CPP approaches in literature are presented in [2], [29] describing their functionalities and applications.…”

“…This algorithm directs the robots to regions that maximize the information gain and distributes swarm robots utilizing an occupancy grid map. Another approach presented in [27] to cover unknown 3D structures utilizing the surface information and centralized multi-robot systems.…”

“…The bounding box of the 3D structure was explored using the multi-robot HCPP approach resulting in a dense point cloud 3D map generated using the collected data at the selected viewpoints in simulation. The dense 3D map could be enhanced further by utilizing the collected data to generate textured 3D map using the approaches described in [27]. The experiments were repeated several times using single and homogeneous multi-robot system of two and three UAVs.…”

Sweep-Your-Map: Efficient Coverage Planning for Aerial Teams in Large-Scale Environments

“…These methods address the coordination problem in frontierbased approaches but suffer from the aforementioned locality problems. The work in [12] extends to the multi-robot case by greedily assigning the view configurations [19]. The work in [20] distributes the workload through continuous region partitioning based on Voronoi components.…”

RACER: Rapid Collaborative Exploration with a Decentralized Multi-UAV System

“…However, such a strategy only allocates a single target to each robot, which may not be effective when multiple robots are assigned to nearby targets. To overcome this limitation, approaches based on segmentation [24,27] and multiple Travelling Salesman Problem (mTSP) [25,26,28] are proposed. Wurm et al [24] divided the already explored area into segments using a Voronoi graph.…”