New visibility-based path-planning approach for covert robotic navigation

Abstract

A new promising approach for visibility-sensitive path-planning problems is presented. The paper focuses on covert navigation where a mobile robot needs to plan a stealthy path to approach a designated destination in a cluttered environment. The aim is to minimize the robot's exposure to hostile sentries within the same environment. The approach can be adapted to work with different levels of initial knowledge the robot may have about both the environment map and the sentries' locations. The approach depends on estimating a cost value at each free-space location that presents the risk of being seen by any sentry. Based on the distance transform algorithm methodology, the minimum visibility–distance cost to a goal is calculated at each cell in the grid-based environment map. Moving along the steepest descent trajectory from any starting point generates an optimal covert path to a goal. The approach has been evaluated with both simulated and physical experiments. A number of test cases are presented. In each case, a path with considerable covertness, compared to a short path to the same destination, is generated. In addition to covert navigation, the approach is introduced briefly as a potential solution for other visibility-based path-planning problems.

Computing highly occluded paths on a terrain

“…Finally, there is some work in GIS and virtual environment communities on finding paths that are occluded from a given set of observers [8,13,14,17]. All these papers sample a set of points on the terrain or polygonal environment, assign a weight to each point based on the visibility from a given set of observers, and use a steepest-descent, A* search, or Dijkstra's algorithm to compute a desired path.…”

Non-Rigid Obstacle Avoidance for Mobile Robots

“…On the other hand, researches on avoidance strategies of non-rigid obstacles scarcely exist except grid-based methods (Marzouqi & Jarvis, 2006;Teng et al, 1993) a few decades ago. These methods were based on the grid map which approximates the environment to spatially sampled space.…”

Occlusion-free path planning with a probabilistic roadmap

“…Marzouqi and Jarvis [7] solve the opposite problem, planning motions to minimize visibility of a robot to hostile observers. Their approach uses a dense 2D grid-based environment map to compute minimum-visiibility paths across a cluttered environment.…”