Search

For many wireless ad-hoc network (WANET) applications, including wireless sensor, robotic, and flying ad-hoc networks, area coverage is a major challenge. This challenge, which may include the number of required nodes, cumulative energy consumption, or total distance travelled, involves covering the largest possible area with the least cost. Whatever the scenario and cost functions, efficiently deploying the nodes throughout their entire missions is an important factor. For larger networks, this can best be performed using proper topology formations that are essentially designated geometric graph patterns. As a network topology formation and node deployment strategy, this study presents a unique tree-formed geometric graph pattern. With any given number of nodes, it guarantees the maximum possible combined area coverage. Considering its graph theory and Euclidean characteristics, the pattern is then described as both algebraic and algorithmic relations. The formation is remarkably scalable because it may include an unlimited number of nodes while ensuring connectivity in a tree topology. In addition to presenting a breakdown of multiple features and attributes, including symmetry and fractality, we will make a comparative analysis with six known regular lattices and a line formation. Our analyses then show that MAX-Tree covers up to twice as much area compared to all other regular lattices of the same node cardinalities while being more robust and reliable than a line formation. MAX-Tree can be extremely beneficial for drone networks, smart industry and agriculture applications, ocean-bed monitoring systems, and many other WANET scenarios.

Ad-hoc networks suffer from dynamicity due to mobility of the nodes or node addition/removal to/from the system. This means, the formed topology may not be connected anymore or better topologies can be found as time passes if the position of the nodes changes or some nodes in the network is broken. Therefore, there is a need for an adaptive solution to make the network’s performance stable against these changes. A centralized manner induces a single point of failure, when the control unit’s service is unavailable, networks can not be formed and communication among the nodes is broken. Hence, a system that distributes the decision-making ability to nodes would be better to increase the robustness of the network. A detailed literature survey about the distributed network formation showed that two critical gaps that need to be filled. Firstly, provided solutions assumed that nodes are using the omnidirectional antennas, hence, they can make transmission and reception in all directions which inherently increases the total interference in the system. Therefore, we come up with the idea of using directional antennas to reduce the interference which requires antennas to point out each other for successful communication. This creates an optimization problem in terms of selecting directions of a certain amount of antennas deployed on a node to maximize the capacity while ensuring connectivity. Secondly, proposed distributed network formation solutions in literature depend on sequential decision-making algorithms that each node waits its turn to make a decision. In that case, deciding or distributing the rank of the nodes is another challenge that needs to be solved by the distributed algorithm. Therefore, in this work, we focused to design a distributed network formation algorithm for a system with directional antennas where all decisions are made simultaneously by nodes in the system. As a result, it is shown that the proposed reinforcement learning-based algorithm is successful and promising to give nodes the ability to adapt to new conditions cooperatively. Ad-hoc-nätverk lider av dynamik på grund av nodernas mobilitet eller nodtillägg/borttagning till/från systemet. Detta betyder att den bildade topologin kanske inte är ansluten längre eller bättre topologier kan hittas allt eftersom nodernas position ändras eller om några noder i nätverket bryts. Därför finns det ett behov av en adaptiv lösning för att göra nätverkets prestanda stabil mot dessa förändringar. Ett centraliserat sätt inducerar en enda felpunkt, när kontrollenhetens tjänst inte är tillgänglig, nätverk kan inte bildas och kommunikationen mellan noderna bryts. Därför skulle ett system som fördelar beslutsförmågan till noder vara bättre för att öka nätverkets robusthet. En detaljerad litteraturundersökning om den distribuerade nätverksbildningen visade att två kritiska luckor måste fyllas. För det första antog förutsatta lösningar att noder använder rundstrålande antenner, och därför kan de göra sändning och mottagning i alla riktningar, vilket i sig ökar den totala störningen i systemet. Därför kommer vi på idén att använda riktade antenner för att minska störningarna som kräver att antenner pekar ut varandra för framgångsrik kommunikation. Detta skapar ett optimeringsproblem när det gäller att välja riktningar för en viss mängd antenner som är utplacerade på en nod för att maximera kapaciteten samtidigt som anslutningsmöjligheterna säkerställs. För det andra är föreslagna lösningar för distribuerad nätverksbildning i litteraturen beroende av sekventiella beslutsfattande algoritmer som varje nod väntar på sin tur för att fatta ett beslut. I så fall är att bestämma eller fördela nodernas rangordning en annan utmaning som måste lösas av den distribuerade algoritmen. Därför fokuserade vi i detta arbete på att designa en distribuerad nätverksbildningsalgoritm för ett system med riktade antenner där alla beslut fattas samtidigt av noder i systemet. Som ett resultat har det visat sig att den föreslagna förstärkningsinlärningsbaserade algoritmen är framgångsrik och lovande för att ge noder förmågan att anpassa sig till nya förhållanden i samarbete.

4th International Conference on Computer Science and Engineering (UBMK) -- SEP 11-15, 2019 -- Samsun, TURKEY Unmanned aerial vehicles (UAVs) are becoming popular in realizing many civilian and military applications. Many conventional military applications require high-cost large and medium-sized UAVs. On the other hands, small UAVs can be adapted to many conventional military applications at a lower cost as compared to large UAVs. In this case, a swarm of small UAVs should carry out the task generally accomplished by large UAVs. Due to the fast movements of the UAVs in a swarm, the communications between the UAVs suffer packet drops due to frequent topology changes. This leads to challenges in maintaining the network topology required for inter-UAV communication. Data loss due to sudden topology changes is especially important for hot-spot areas generally located at the center of the topology. In this study, a balanced topology formation approach is proposed by taking the data traffic and node density of the UAV network into account. Furthermore, the network is divided into clusters with the aim of optimising the network delay and reducing packet drop rate. The results obtained from the clustering algorithm are compared with the previously proposed methods for UAV and mobile ad-hoc networks (MANET). The results have shown that the proposed method balances data traffic better and a reduced packet loss rate is achieved as compared to the previously proposed methods. IEEE, IEEE Turkey Sect WOS:000609879900133 2-s2.0-85076223154

This paper presents a novel algorithm for clustering of nodes by transmission range based clustering (TRBC).This algorithm does topology management by the usage of coverage area of each node and power management based on mean transmission power within the context of wireless ad-hoc networks. By reducing the transmission range of the nodes, energy consumed by each node is decreased and topology is formed. A new algorithm is formulated that helps in reducing the system power consumption and prolonging the battery life of mobile nodes. Formation of cluster and selection of optimal cluster head and thus forming the optimal cluster taking weighted metrics like battery life, distance, position and mobility is done based on the factors such as node density, coverage area, contention index, required and current node degree of the nodes in the clusters

Users of device-to-device (D2D) communication need efficient content discovery mechanisms to steer their requests toward the node in their neighborhood that is most likely to satisfy them. The problem is further compounded by the lack of a central coordination entity as well as by the inherent mobility of devices, which leads to volatile topologies. In this paper, we first discuss group-based communication among non-rooted Android devices using Wi-Fi direct, a protocol recently standardized by the Wi-Fi alliance. We propose intra- and inter-group communication methodologies, which we validate through a simple testbed where content-centric routing is used. Next, we address the autonomous formation of groups with the goal of achieving efficient device resource utilization as well as full connectivity. Finally, we evaluate the performance of our group formation procedure both in simulation and in a real testbed involving Android devices in different topologies.

Micro and nanorobotics represents one of the most challenging sectors of modern robotics. Through batch fabrication of Micro Electro-Mechanical Systems (MEMS), advanced small scale sensing and actuating tasks in a wide area of applications can be performed. Most miniaturized electro-mechanical devices are characterized by low-power and low-memory capacity. The huge number of modular robots introduces the need to explore novel self-reconfiguration algorithms to optimize movement and communication performances in terms of efficiency, parallelism and scalability. Nano-transceivers and nano-antennas operating in the Terahertz Band are already a well acquainted communication paradigm, enforcing nano-wireless networking that can be directly integrated in MEMS microrobots. Several logical topology shape-shifting algorithms are already implemented and tested in literature, along with performance evaluation on nano-wireless use. This article aims to provide an algorithm to reconnect groups of microrobots, along with a novel movement model for microrobotics ensembles introduced to enforce more realistic simulations. Special emphasis is given on the need of novel movement algorithms for swarms of microrobots.

Among all the new robotics technologies, great interest is rising in the field of microrobotics. Micro Electro-Mechanical Systems (MEMS) microrobots are miniaturized electro-mechanical devices that can perform various tasks in a wide area of applications. Despite the low-power and low-memory capacity, they are provided with sensors and actuators. Self-reconfiguration is a key factor for MEMS microrobots to perform their tasks and to optimize their communications in order to achieve efficiency, parallelism and scalability. Nano-transceivers and nano-antennas operating in the Terahertz Band offer a promising communication paradigm, providing nanowireless networking directly integrated in MEMS microrobots. Catoms from Claytronics project are an appropriate microrobotics case to explore this novel framework. Several logical topology shape-shifting algorithms have been implemented and tested, along with different nano-wireless simulations. This paper aims to provide a survey on nano-wireless communication for modular robotics and propose some optimization choices. Special emphasis is given to the use of the nano-wireless communications for topology formation and maintenance in microrobotics.

Wireless Sensor Networks are traditionally composed of a multiplicity of sensor nodes that sense given phenomena and deliver the sensed data to specific sink nodes. In the most of the application scenarios, sensor nodes have been considered motionless. On the contrary, interesting possibilities arise if some sensors are embedded in devices carried by mobile agents as people, cars, animals, etc. If sinks move within the considered sensor field, they are able to provide both sparse sensing and collecting of data measured by static sensors placed at fixed locations. The main goal of this work is to evaluate, through simulations, the impact of sinks' mobility in a wireless sensor network created by using the topology formation mechanism provided by the IEEE 802.15.4 Standard. To this aim, as a practical case study, we consider a wireless sensor network deployed in a museum used to monitor the presence, the localization and other parameters of artworks exposed in it. In this context, we analyze how sinks' mobility affects connectivity and energy consumption for network formation and re-configuration.

The IEEE 802.15.4 standard defines a set of procedures to set-up a low-rate wireless personal area network where nodes self-organize into a logical communication structure through which data can be routed, hop by hop, from sources to destinations. The network formation of the IEEE 802.15.4 does not impose constraints on the topology. The Zigbee alliance uses the IEEE 802.15.4 layers to build a complete protocol stack for the implementation of wireless sensor networks. ZigBee specifies the network layer for star, tree and peer-to-peer topologies. Starting from these, more complex cluster-tree topologies can be formed. To control the network topology ZigBee fixes the maximum number of routers and end-devices that each router may have as children and also fixes the maximum depth of the tree. To better understand the importance of these constraints we simulate and analyze the IEEE 802.15.4 formation procedure in different network settings (single-sink and multi- sink scenarios). The goal is to provide guidelines for the practical implementation of ZigBee network formation with the aforementioned constraints.

This paper investigates the topology formation mechanisms provided by IEEE 802.15.4 to create tree-based topologies, in the context of a multi-sink wireless sensor network. the topology formation protocol is simulated starting from a random deployment of a given number of nodes, to obtain statistics of the position of those nodes belonging to given tree levels. It is shown that under some circumstances the nodes belonging to any level can be assumed to be uniformly distributed over the plane. Under such assumption, a mathematical formulation is used to optimize the average number of children per parent and the number of levels in the tree (that is, the tree height), through maximization of the network association probability