Establishing Relay Network Communication for Mobile Units Via Autonomous Vehicles

People

Md Mahbubur Rahman, Leonardo Bobadilla, Brian Rapp, Franklin Abodo, Gregory Reis, Virgil Acuna

Abstract

In communication-denied or contested environments, Line-of-Sight(LoS) communication (by infrared, visible light, or long distance RF) becomes the most reliable and efficient way to send information between geographically scattered mobile units. In this paper, we consider the problem of planning trajectories and locations for a group of autonomous vehicles to serve a set of computational light units that are dispersed in an environment with obstacles. The vehicles should find trajectories that will allow them to cover all units and additionally form a relay network among the vehicles. The contribution of the paper is the following: 1) We proposed centralized and distributed algorithms to verify that the vehicles and unit form a connected network through LoS; 2) We proposed an algorithm that can maintain connectivity, if possible, by relocating a single vehicle; 3) We studied the computational complexity of this class of problem in multi vehicle movement, proved that it is NP-Hard, and proposed approximated procedures to calculate suitable paths for the vehicles. We tested our ideas through 1) Algorithmic implementation; 2) A case study in the form of a computer simulation and 3) Hardware design and deployment.

Motivation

In a communication denied environment with limited transmission capability, Line of Sight cmmunication is the most effective and trusted way for service exchange. This types of networks are difficult for a jammaer or attacker to intercept. Here we develop a methodology to deploy a number of autonomous units in a communication denied environment due to obstacles and abrupt weather. Visibility based Line of Sight communication is considered as a medium of information exchange. In order to establish such network we depoly a number of servicing vehicles equipped with high performance computing module to serve the mobile units. Then a valid network is established with the following properties:

A deployed unit needs to be inside the visibility polygon of at least one vechile to receive service.
A vehicle needs to be in the visibility region of at least one other vehicle to form a connected relay network.

In case we cannot establish a static relay network due to lack of servicing vehicle, we propose an approximate solution inspired by the set cover problem that identifies a number of suboptimal goal regions for the vehicles. A TSP tour is proposed for traversal of the vehicles along the regions.

Details

The below images and videos demonstrate, using a physical environment and a virtual simulator, how a network status is calculated that is composed of distributions of autonomous vehicles and mobile units. Moreover a proof of concept is presented for the proposed methodology to calculate the motion paths of the vehicles in order to keep all the units connected. We used different open source libraries for siulation and hardware deployment.

Software Tools: Robotic Operating System, Visilibity Library, OMPL, CGAL, Rviz, Gazebo, GraphViz and OpenCV.
Hardware Tools: Traxxas Truck, APM flight controller, GPS module, Raspberry pi, Zigbee, Arduino, Turnigy radio control.

Visibility Polygon: We connected the Visilibity library with ROS node to calculate the visibility polygons of all the units.
Graph Search: BFS search algorithm was used and the visited nodes were colored using gray. Any white node indicates invalidity of the system.
Gazebo Environent: A vehicle was used to simulate motion using ROS and OMPL. The visualization was sent to the Gazebo node from ROS.
Hardware Deployment: We designed a vehicle as a servicing module by modifying a Traxxas truck. We mounted Raspberry pi and connected a camera to it for visibility purposes. Also zigbee communication system was used along with Arduino board for wireless communication. GPS module was used with an APM flight controller to get the position of the vehicle and calculate the waypoints to follow.