Perception-aware Exploration for Consumer-grade UAVs

TL;DR

The paper tackles autonomous exploration and mapping using consumer-grade UAVs, addressing limitations in sensing, computation, and communication. It combines a depth-estimation-aware viewpoint-pair sampling method with a yaw-aware trajectory planner and a semi-distributed base-node coordination to extend single-UAV exploration to multiple drones. Key contributions include sampling and evaluating viewpoint pairs for depth reconstruction, adapting the hierarchical planner to pair-based viewpoints, introducing a yaw-constrained trajectory optimization, and implementing a semi-distributed multi-UAV framework that outperforms a baseline across multiple agents. Experiments in simulation demonstrate improvements in depth accuracy, map quality, and odometry stability, validating feasibility for consumer UAVs and showing scalable performance with more aerial agents.

Abstract

In our work, we extend the current state-of-the-art approach for autonomous multi-UAV exploration to consumer-level UAVs, such as the DJI Mini 3 Pro. We propose a pipeline that selects viewpoint pairs from which the depth can be estimated and plans the trajectory that satisfies motion constraints necessary for odometry estimation. For the multi-UAV exploration, we propose a semi-distributed communication scheme that distributes the workload in a balanced manner. We evaluate our model performance in simulation for different numbers of UAVs and prove its ability to safely explore the environment and reconstruct the map even with the hardware limitations of consumer-grade UAVs.

Figures (4)

• Figure 1: Overview of the proposed pipeline for single-UAV exploration.
• Figure 2: Map for different viewpoint pairs evaluation methods.
• Figure 3: Map for trajectory generation methods.
• Figure 4: Map and path for semi-distributed exploration method and different numbers of UAVs.