Achieving quality of service in wireless sensor networks using cellular learning automata

Computer-Software Master Thesis (M.Sc)

Abstract

The quality of service in wireless sensor networks is very different compared to traditional networks. Some of the parameters that are used in evaluating the quality of service in these networks are: network coverage, optimal number of active nodes in the network, network lifetime and energy consumption. In this thesis, three basic problems of wireless sensor networks have been raised and with the aim of improving service quality parameters, effective solutions have been presented for these problems using the intelligent method of cellular learning automata. First, the problem of environment coverage in sensor networks is solved by deactivating the unnecessary nodes and keeping the nodes optimally active, in order to save energy and increase the life of the network. Then the problem of clustering in the sensor network is addressed and by using cellular learning automata, the sensor networks are clustered in such a way that the energy is uniformly consumed in the network and the life of the network is increased. After that, using learning automata, a method of aggregating data from the sensor environment is proposed, which saves energy in the network and increases the life of the network. All presented methods have been simulated using J-Sim software. The results of the simulations show the better performance of the proposed methods compared to similar methods. Keywords: wireless sensor networks, learning automata, quality of service, coverage, clustering, data aggregation 1- Introduction Wireless sensor networks Wireless sensor networks SIM[1] are used to collect information in areas where the user cannot be present. In a sensor network, sensors individually sample (measure) local values ??and send this information to other sensors if necessary and finally to the main observer. The function of the network is to report the phenomena that are happening to the observer who does not need to know anything about the structure of the network and the sensors individually and their relationship. These networks are independent and self-governing and work without human intervention. Usually, all the nodes are identical and practically by cooperating with each other, they fulfill the overall goal of the network. The main goal in wireless sensor networks is to monitor and control atmospheric, physical or chemical conditions and changes in an environment with a certain range [1, 2]. Wireless sensor network is a special type of ad hoc networks [2].  The topic of wireless sensor networks is one of the new topics in the field of network engineering and information technology.

Recent advances in the design and manufacture of commercial chips have made it possible to perform signal processing and sensing in a single chip, a wireless network sensor, which includes microelectromechanical systems [3] (MEMS) such as sensors, actuators [4] and RF radio components.

Small wireless sensors have been produced that can be assembled It has the ability to collect data from a distance of several hundred meters and send data between wireless sensors to the main center, and with this technology information on temperature - fluctuations, sound, light, humidity, and magnetism can be collected. These wireless sensors can be installed in wireless sensor networks with low cost and small size. But the miniaturization of wireless sensors also has disadvantages. Semiconductor technology has led to the creation of fast processors with high memory, but powering these circuits is still a major problem, which is limited to the use of batteries. The power supply section is an important and limited section that if batteries are used in these networks, replacing the batteries in the case of a large number of network nodes will be a difficult task, and the nodes will be forced to use short-range communications in order to save and save energy. The difference between an efficient wireless sensor and an energy-inefficient wireless sensor is in their performance in hours versus weeks. Increasing the size of the WSN network causes the complexity of routing and sending information to the main center. But routing and processing still require energy. Therefore, one of the key points in the development and presentation of new routing algorithms is to reduce and save energy consumption. Different parts of wireless sensor networks should be simulated and modeled to evaluate their effectiveness.For this, wireless sensor networks are mapped to graphs in which each node corresponds to a node in the network and each edge represents a link or communication channel between two nodes in the network. If the connection between the nodes in the network is bidirectional, the mapped graph will be undirected, and if the connection between the nodes in the network is asymmetric, then the mapped graph will be directional. Of course, the communication model between nodes in the network can be one-to-one or one-to-all. Providing a practical model for sensors is a complex and difficult task, which is due to the variety of sensors both in terms of structure and in terms of the principles and basis of their work. Sensor networks have unique features, which has caused special protocols to be considered for them.

In wireless sensor networks, there are usually only one or two base stations and a large number of sensor nodes are distributed in the environment. Due to the limited range of these sensors and battery energy, many nodes are not able to communicate directly with the base station. But by relying on nodes like itself and other sensor nodes, it communicates with the base station, which in MANET networks [5] this action is also done by normal nodes. In wireless sensor networks, a large number of nodes with communication, processing, environment sensing and They are scattered in an environment with a certain framework. The event that happened or the questions asked by the central node[6] and the mission assigned to each node cause communication between the nodes. The exchanged information can be a report of the state of the area under the control of the sensor nodes to the central node or a request from the central node to the sensor nodes. The central node, as the communication port of the sensor network with other systems and telecommunication networks, is actually the final receiver of the report from the sensor nodes and after performing a series of processes, it sends the processed information to the user (using a communication medium such as the Internet, satellite, etc.). On the other hand, user requests are also transmitted to the network by this node.

A sensor node can assume one of the two roles of producing data or relaying data produced by other nodes. Generally, in sensor networks, most nodes play both roles together. Establishing and designing the communication structure and architecture between network nodes requires compliance with many different factors, including fault tolerance, scalability, production cost, operation environment, sensor network topology, hardware limitations, communication tools and media, energy consumption, etc. is To learn more about wireless sensor networks, refer to the first appendix. 2-1-1 Issues in wireless sensor networks Several factors are effective in the design of sensor networks, and there are many issues in this field that cannot be reviewed in this article, so we will only briefly mention some of them. 1- Routing: The main nature of sensor networks is that the tasks they perform It must be local because each node can only communicate with its neighbors and general and global information from the network is not very available (collecting this information consumes a lot of money and time). The information obtained by the nodes must be sent to the central node using routing techniques.

2- Hardware bottlenecks: each node should have all the necessary components and should be sufficiently small, light and compact. At the same time, each node should have very low energy consumption and low cost and be compatible with environmental conditions. These are all limitations that challenge the design and construction of sensor nodes. Providing lightweight and compact hardware designs for each of the node components, especially the wireless communication and sensors, is one of the research topics that has a lot of work to do. The technological progress of manufacturing integrated circuits with high compression and low consumption has played a significant role in reducing hardware bottlenecks.

3- Fault tolerance and reliability[7]: Each node may be damaged or destroyed by environmental events such as accidents or explosions, or fail due to the end of the energy source. Tolerability or reliability means that the failure of nodes should not affect the overall performance of the network.