Flocking and topology manipulation based on space partitioning

doi:10.1016/j.robot.2019.103328

Robotics and Autonomous Systems

Volume 124, February 2020, 103328

https://doi.org/10.1016/j.robot.2019.103328 Get rights and content

Highlights

•
Topology connectivity defined by enclosed areas from overlapping communications ranges.
•
An algorithm to determine the degree of similarity between topologies.
•
Independent and coordinated decision-making based on topology similarity and enclosed area.
•
Guidance for vehicles to enhance network robustness against formation perturbations.

Abstract

Network topology plays a critical role in enabling a multi-agent system to adapt to environment changes and achieve desired objectives. This paper presents distributed topology manipulation schemes for a group of mobile agents. The agents have limited heterogeneous communication ranges, and connections among them are directional. The topology is established from the overlapping communication ranges. The admissible space is partitioned into enclosed areas by connectivity among the agents based on their communication ranges. Each agent occupies an enclosed area, and its decision-making manipulates the topology by guiding itself to an adjacent enclosed area. Both independent and coordinated decision-making approaches are provided. A guidance algorithm is designed to drive the vehicles to a flexible formation, in which the robustness of the network topology is enhanced.

Introduction

Many environmental factors and energy restrictions limit radio communications in terms of range and capacity. To address the challenges, range extension [1] and data ferrying [2], [3] mechanisms have been proposed with UxVs (unmanned aerial and ground vehicles) being used to assist communications. For range extension of the field autonomous network (OPAL) in [1], the UAVs use formations to maintain a network. However, there is no guarantee that communication among the unmanned resources will remain whilst the resources are traversing the network. Similarly data ferrying approaches investigate the possibility of using single data ferries via scheduling or route planning, and in terms of data ferrying, the network is disconnected. Our work examines the use of coordinated unmanned agents which work as a single entity in terms of one flock/fleet of vehicles.

Research works have addressed similar problems using free space optical links [4], requiring specialized hardware for directional communication. Others have used flocking based on attraction and repulsion rules in [5], and examined the connectivity of a flock. However, it is not meant for enabling connectivity with others and it does not provide the ability to maintain flock connectivity when faced with obstacles. Simple unmanned chaining mechanisms in [6] and data ferrying based chaining mechanisms in [7] have been proposed, but these approaches do not allow adaptability in shapes, in which the fleet of vehicles can take. The approach proposed in this paper seeks to ensure that the unmanned fleet armed with omni-directional antennas is enabled with topology manipulation. We envisage this mechanism as being able to provide an adaptive RF footprint to perform data-ferrying with reduced movement of the flock as a whole, and to carry out range extension functionality. Such an approach has applicability in the coordination of swarm drones for hazard awareness and providing network service.

Communication topology plays a vital role in network systems [8], [9], [10]. In a network of mobile vehicles, the limited and different ranges of communication among the vehicles makes maintaining connectivity challenging. The formation has to be flexible enough to cope with complicated environments and handle challenging tasks [11], [12], [13]. Our goal is to achieve a specific network topology. The desired network topology can be achieved by multiple formations. In our work, each formation that leads to the desired network topology is equally desirable, and the ability to achieve multiple formations demonstrates the flexibility of the algorithm. Our solutions are based on space partitioning and enclosed areas. The use of enclosed areas enables the vehicles to reach a desired network topology in flexible formations, and the capability of self-organizing allows the vehicles to operate in complicated circumstances.

Vehicles in the form of robots with exchangeable positions were studied in [14], [15], [16], [17], [18], [19]. A recurrent neural network was used in [16] to find the optimal formation. Such a method could be applied to a large-scale system due to its capacity of parallel computation. Simulated annealing, genetic algorithm and particle swarm optimization were tested in [14], and a team of wheeled robots were assigned to deliver an object along a certain path. Multiple leader candidate approach was also used to find the most suitable leader in [15]. Such an on-line method was robust against possible failures of leaders and followers. In driving the robots to the desired formation, a predefined network graph is assumed in [20]. The role of the controller is to guarantee the graph connectivity based on the connectivity weights from the neighborhood.

Flexible formations were investigated in [21], [22], [23]. Shape orientated formation control is adopted in [21]. Non-uniform rational basis spline is used to produce mapping for the individual robots, and weighted control points and constraints are inputs for such representation such that the nominal curve could be globally smooth. Coalition formation problem is regarded as a multi-objective optimization problem in [23], and parallelized algorithms are designed to address the issue of scalability. Cooperative task scheme is also used in deriving the optimal formation in [22], but in combat with parametric uncertainty, a global control algorithm is designed.

In this study, we address the topology manipulation problems for coordinating unmanned vehicles to work as a single entity in terms of one connected flock of vehicles based on enclosed areas from space partitioning. Our contributions are listed as follows.

•
A concept of topology connectivity within a formation defined by enclosed areas from space partitioning of overlapping communications ranges;
•
An algorithm to determine the degree of similarity between topologies;
•
Decision-making algorithms based on topology similarity and enclosed area. They include independent decision-making and coordinated decision-making;
•
Guidance design for the vehicles to enhance network robustness against formation perturbations.

This paper begins with Section 2 Problem Statement. Section 3 introduces a novel representation of connectivity based on enclosed areas, and describes two space partitioning methods to derive enclosed areas. In Section 4 network topology is explained from the perspective of data paths. Section 4 gives topology manipulation methods under fixed mapping from vehicle locations to network nodes. The decision-making is made based on similarity evaluation between different network topologies. Section 6 derives further solution for topology manipulation in the search of the optimal mapping. Solutions for both independent and coordinated vehicles are provided. In Section 7, a guidance algorithm is designed to enhance the robustness of current network topology. In Section 8, the proposed methods are demonstrated and their performance is evaluated via simulation studies of topology manipulation of a vehicular network of five vehicles. The paper is concluded in Section 9.

Section snippets

Problem statement

Network topology and vehicle formation are studied in this paper. It is assumed that the vehicles have limited communication ranges. Vehicles receive data from other vehicles that are inside the communication ranges. Omnidirectional antennas are assumed and heterogeneous ranges exist within the network, creating a directed vehicular network.

Given a vehicle formation in space, we obtain a network under a given set of communication ranges. The network topology changes if the formation is changed.

Space partitioning and enclosed area

This section introduces the concept of enclosed areas. To derive the enclosed areas, two space partitioning methods are proposed.

Network topology based on data path

A network topology can be specified with the incoming data paths/routes, into the individual nodes from the rest of the nodes in the network. The vehicular network and the desired network can be described in similar ways. For example, there are $n$ nodes in desired network $N$ , that is $N = \{\begin{matrix} N_{1} & N_{2} & \dots & N_{n} \end{matrix}\}$ .

For node $N^{'}$ , the data path $Ψ (N^{'})$ to $N^{'}$ is defined: $Ψ (N^{'}) = \{\begin{matrix} \dots & N_{2}^{'} & N_{1}^{'} \end{matrix}\},$ where node $N^{'}$ receives data from $N_{1}^{'}$ , node $N_{1}^{'}$ receives data from $N_{2}^{'}$ , $\dots$ .

Suppose that for node $N_{1}$ , its incoming data paths are from nodes $N_{2}$ ,

Topology manipulation under fixed mapping

In this section, the aim is to drive the vehicles under a fixed mapping such that the network topology of $M$ is the same as the network topology of $N$ .

Topology manipulation for optimal mapping

In the previous section, algorithms for decision-making and guidance are designed to drive the individual vehicles to the desired network under a fixed mapping. However, this means that a fixed mapping is needed and such fixed mapping is not likely to be capable to handle a wide range of circumstances. In this section, we design an adaptive mapping scheme such that the vehicles are able to find the local optimal mapping with smallest number of topology mismatches.

Topology robustness of network

The guidance algorithm in (9) drives the vehicles to their desired enclosed area. However, there are significant uncertainty in vehicle modeling and guidance, and detrimental vibrations in actuators and external disturbances are impossible to eliminate. Consequently, there is no way to predict the precise positions of the vehicles. The jittering vehicles create jittering arcs on their communication ranges. This results in jittering arcs in all the enclosed areas. For the vehicles that are close

Simulations

In this section, we present simulations to display and contrast the output of the algorithms on an example network topology. Simulations are run for topology manipulation of a vehicular network $M$ with 5 vehicles, under fixed mapping, independent mapping, and coordinated mapping, and guidance algorithm (9). The software is Matlab 2015b on OS X system with 2.5 GHz Intel Core i7 processor. The initial positions of the vehicular network $M$ in 2D coordinate are $M_{1} = [\begin{matrix} 672 \\ 574 \end{matrix}], M_{2} = [\begin{matrix} 445 \\ 635 \end{matrix}], M_{3} = [\begin{matrix} 652 \\ 454 \end{matrix}],$ $M_{4} = [\begin{matrix} 318 \\ 430 \end{matrix}],$

Conclusion

This paper investigates the topology manipulation problems among independent vehicles and among coordinated vehicles based on the concept of enclosed areas. The algorithms of decision-making and guidance presented are able to drive network topologies from a current vehicular network to the desired network using mapping functions. A function to evaluate the similarity between networks is provided. Decision-making of the individual vehicles is distributed, and the vehicles receive information

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Dr Hongjun Yu received his PhD degree from the University of Adelaide. He worked as research assistant and postdoctoral fellow in the University of Adelaide until July 2018. He is now working as a lecturer in Harbin Engineering University.

References (26)

LlorcaJ. et al.
A quadratic optimization method for connectivity and coverage control in backbone-based wireless networks
Ad Hoc Netw.
(2009)
AnejaY.P. et al.
A branch-and-cut algorithm for the strong minimum energy topology in wireless sensor networks
European J. Oper. Res.
(2010)
JanssenR.H.P. et al.
Stable network topologies using the notion of covering
European J. Oper. Res.
(2012)
MonsuurH.
Stable and emergent network topologies: A structural approach
European J. Oper. Res.
(2007)
AbbaspourA. et al.
Optimal formation and control of cooperative wheeled mobile robots
C. R. Mec.
(2015)
MengY. et al.
A morphogenetic approach to flexible and robust shape formation for swarm robotic systems
Robot. Auton. Syst.
(2013)
DydekZ.T. et al.
Adaptive configuration control of multiple UAVs
Control Eng. Pract.
(2013)
AgarwalM. et al.
Parallel multi-objective multi-robot coalition formation
Expert Syst. Appl.
(2015)
HuiK.P. et al.
OPAL-A Survivability-oriented approach to management of tactical military networks
UsbeckK. et al.
Data ferrying to the tactical edge: A field experiment in exchanging mission plans and intelligence in austere environments

HunjetR. et al.

Survivable communications and autonomous delivery service a generic swarming framework enabling communications in contested environments

SaberR.O. et al.

Flocking with obstacle avoidance: Cooperation with limited communication in mobile networks

DixonC. et al.

Maintaining optimal communication chains in robotic sensor networks using mobility control

Mob. Netw. Appl.

(2009)

Cited by (5)

Model-based learning of underwater acoustic communication performance for marine robots
2021, Robotics and Autonomous Systems
Accurate prediction of acoustic communication performance is an important capability for marine robots. In this paper, we propose a model-based learning methodology for the prediction of underwater acoustic communication performance. The learning algorithm consists of two steps: (i) estimation of the covariance matrix by evaluating candidate functions with estimated parameters using detrended measurements;and (ii) prediction of communication performance. Covariance estimation is addressed with a multi-stage iterative training method that produces unbiased and robust results with nested models. The efficiency of the framework is validated with simulations and experimental data from field trials. The field trials involved a manned surface vehicle and an autonomous underwater vehicle.
Stochastic control and time scheduling for irregular robots
2021, Journal of the Franklin Institute
Citation Excerpt :
In this paper, we propose control design that reduces computation load and requires only simple rules without solving complicated equations. Existing literature on formation control has focused on model based approach [19,28], hierarchal structure [3,5,9,11] and objective function [1,7,25,26]. Finite control set is used to simplify control design in [19] and deformation based on sub destination is used as reference.
Vehicles of different sizes are difficult to navigate in close vicinity. In this paper, we propose multi-vehicle coordination strategy by stochastic control and time scheduling to guarantee no collisions. We use contours and relative motions of vehicles to calculate collision time and use it in multi-vehicle scheduling and reduce computation burden. The proposed strategy enables the vehicles to add calculation delay when vehicles are moving towards the destinations. To avoid complicated control rule design for motion-restricted irregular vehicles, we propose stochastic control to provide satisfactory performance. By changing the frequency of control update, a modification is proposed to take congestion into account. Simulation examples are given to demonstrate the effectiveness of the proposed approach.
Hierarchical Leader-Follower based Regular Topology Formation Control Scheme in 5G-enabled Distributed Wireless Robotic Networks
2022, ACM International Conference Proceeding Series
Steering Control and Path Planning of Two-Wheel Vehicle with Hazard Avoidance
2021, Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics
A Survey on Intelligent Control for Multiagent Systems
2021, IEEE Transactions on Systems, Man, and Cybernetics: Systems

Professor Cheng-Chew Lim received his PhD degree from Loughborough University. He worked as a lecturer at the National University of Singapore from 1981 to 1987. He has been with the University of Adelaide since 1987, and is currently a professor and Director of Research in the School of Electrical and Electronic Engineering.

Dr Robert Hunjet has worked for Australia’s Defence Science and Technology Group since 2001. He is the Group Leader of their Advanced Vehicle Systems Science and Technology Capability and is an adjunct Associate Professor at UNSW Canberra’s School of Engineering and Information Technology. He received his PhD in 2012 from the University of Adelaide for his thesis on the use of autonomy to enable adaptive network topologies. Robert serves as the co-lead for the Networked Land Autonomy theme within the Defence Collaborative Research Centre on Trusted Autonomous Systems. He leads a research team which seeks to enable autonomy through distributed control of sensors and effectors across federated land vehicles. His research interests are wireless network performance, swarming and emergence, autonomous decision making, network survivability and distributed control.

Professor Peng Shi was awarded the Doctor of Science degree by the University of Glamorgan, UK in 2006, and the Doctor of Engineering degree by the University of Adelaide, Australia in 2015. He is now a professor in the University of Adelaide. He was a chair professor in the University of Glamorgan, UK, and a senior scientist in the Defence Science and Technology Organisation, Australia. He was the Chair of Control Aerospace and Electronic Systems Chapter, IEEE South Australia Section, and is a Fellow of IEEE, IET, IEAust.

View full text

Flocking and topology manipulation based on space partitioning

Highlights

Abstract

Introduction

Section snippets

Problem statement

Space partitioning and enclosed area

Network topology based on data path

Topology manipulation under fixed mapping

Topology manipulation for optimal mapping

Topology robustness of network

Simulations

Conclusion

Declaration of Competing Interest

Ad Hoc Netw.

European J. Oper. Res.

European J. Oper. Res.

European J. Oper. Res.

C. R. Mec.

Robot. Auton. Syst.

Control Eng. Pract.

Expert Syst. Appl.

OPAL-A Survivability-oriented approach to management of tactical military networks

Data ferrying to the tactical edge: A field experiment in exchanging mission plans and intelligence in austere environments

Survivable communications and autonomous delivery service a generic swarming framework enabling communications in contested environments

Flocking with obstacle avoidance: Cooperation with limited communication in mobile networks

Maintaining optimal communication chains in robotic sensor networks using mobility control

Mob. Netw. Appl.