Visibility-Aware Navigation With Batch Projection Augmented Cross-Entropy Method Over a Learned Occlusion Cost

Abstract

We present two real-time trajectory optimizers based on the Cross-Entropy Method for visibility-aware navigation. The two approaches differ in handling inequality constraints stemming from bounds on motion derivatives, collision avoidance, tracking error, etc. Our first optimizer augments the inequalities into the cost function, while the second one relies on a novel GPU accelerated batch projection algorithm. We adopt a learningbased approach to ensure a fast query of the occlusion cost arising from the environment. Specifically, we train a neural network to compute the occlusion directly from the point obstacles generated from LiDAR or RGB-D sensors. Our learned occlusion model can be queried up to 3x faster than the approaches based on distance computation from occupancy or voxel maps. We improve the state-of-the-art in the following aspects. First, our optimizers do not require any explicit map building and can thus adapt on the fly to the changes in the environment. Second, we outperform existing approaches in target tracking applications in maintaining target visibility and success rate while being competitive in acceleration effort and computation time.

PRIEST: Projection Guided Sampling-Based Optimization for Autonomous Navigation

“…In contrast, increasing the covariance of control perturbation has been shown to make MPPI diverge [11]. Improvement over Author's Prior Work [15] PRIEST builds on our prior work [15] that used projection-augmented sampling for visibility-aware navigation. Our current work targets a much broader scope of navigation problems with potentially non-smooth costs.…”

ROBOTONT – Open-source and ROS-supported omnidirectional mobile robot for education and research

“…Robotont is being continuously validated as an educational tool and a research platform - notable examples of such results have already been reported in a number of research publications [20] , [21] , [22] , [23] , [24] , [25] , [26] , [27] . More specifically, this robot has been extensively used to deliver university teaching, professional training, and online courses on the topics of ROS and robotics as described in [20] , [21] , [22] , [23] .…”

“…[28] , [29] , [30] , [31] , [32] ) and a platform to validate cutting-edge robotics research about e.g. software frameworks [24] and motion-planning [25] , [26] , [27] .…”

Cooperative Navigation and Guidance of a Micro-Scale Aerial Vehicle by an Accompanying UAV using 3D LiDAR Relative Localization

“…In [19], the UAV shared its voxel map of the environment with a UGV, which converted the voxel map to an elevation map and utilized it for path planning. In [20], a trajectory optimization approach for a UGV guiding a UAV while maintaining visibility was proposed, motivated by scanning warehouse racks for inventory management.…”