Minimal Exposure Escape Path Problem

doi:10.1109/TII.2023.3336354

Minimal Exposure Escape Path Problem

2024-04-01

0

PRE

AI

A

Armando Alves Neto *

V

Víctor Costa da Silva Campos

D

Douglas G. Macharet

DOI：10.1109/TII.2023.3336354

Original

Original request for help

Save

Abstract

En 中文

This article introduces the minimal exposure escape path (MEEP) problem, in which a mobile robot must determine a path from a start position to a safe region that less exposes it to any sort of industrial environmental risks. It is a generalization of the classic minimal exposure path (MEP) problem, where the exposure is associated with the detection by an industrial wireless sensor network (IWSN). The MEEP problem is applicable in many situations when it is necessary to escape from or penetrate predefined zones avoiding threats along the way, a common task in military operations, or search and rescue missions. We model it as an optimal control problem and use a semi-Lagrangian numerical scheme to solve it. We show that our approach asymptotically converges to the optimal solution while still allowing us to incorporate different characteristics and typical real-world constraints. Simulated experiments show the applicability of our method in 2-D and 3-D spaces, considering obstacle-free or cluttered environments, with nonconvex and multiple goal regions, and heterogeneous networks.

Keywords:

Dynamic programming

industrial wireless sensor networks (IWSNs)

minimal exposure path (MEP)

policy iteration

Minimal Exposure Escape Path Problem

Abstract

Journal

Organization