RETT-gen: A globally efficient routing protocol for wireless sensor networks by equalising sensor energy and avoiding energy holes
abstract:
Wireless sensor networks are composed of a large number of sensor nodes with limited energy resources. Once deployed, the sensor nodes are usually inaccessible to the user, and thus replacement of the energy resource is not feasible. Hence, energy efficiency is a key design issue that needs to be enhanced in order to improve the life span of the entire network. Several routing protocols have been proposed to improve the effective lifetime of the network with limited energy supply. In this paper, we propose routing based on energy–temperature transformation, RETT-gen, a scalable energy-efficient clustering and routing protocol designed for wireless sensor networks. The main goal of RETT-gen is to evenly distribute the energy load among all the sensor nodes in the network so that there are no overly-utilized sensor nodes that will run out of energy before the others. To achieve this goal, RETT-gen uses heat conductivity as a metaphor and uses the heat dissipation difference equations. In RETT-gen, we transform the expected lifetime of each sensor node to an equivalent temperature, and then by using the heat dissipation equations, we find the hottest path for sending data to the base station, which will not always be the shortest path. We evaluate the performance of the RETT-gen protocol via simulations, and compare it to the performance of well-known routing protocols (i.e. LEACH [W. Heinzelman, A. Chandrakasan, H. Balakrishnan, Energy-efficient communication protocol for wireless microsensor networks, in: Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS’00), 2000.] and EEUC [C. Li, M. Ye, G. Chen, J. Wu, An energy-efficient unequal clustering mechanism for wireless sensor networks, in: Proceedings of the International IEEE Conference on Mobile Adhoc and Sensor Systems (MASS), 2005.]). Simulation results show that by equalizing the sensor nodes energy, RETT-gen insures that the lifetime of the entire sensor network is maximized, the connectivity in a sensor network is maintained for as long as possible, and that the residual energy of the entire network is of the same order.
Introduction
Wireless sensor networks consist of a large number of densely deployed sensor nodes. The sensor nodes normally feature a low-power wireless radio, an embedded processor, some sensors (i.e., temperature, light, acceleration, etc.) and a small battery. The sensor nodes send their sensed data, usually via radio, to a base station either directly or relayed via other sensor nodes. In order to maximize the lifetime of a sensor network, efficient energy routes need to be selected for the sensor nodes in a reliable manner while operating under strict constraints in computation, communication and energy resources. These unique features have raised some interesting problems that must be addressed when designing a routing protocol for sensor networks. The protocol must be energy efficient, scalable, and adaptable. Moreover, the energy resources of each individual sensor node must be managed effectively by the protocol.
In this work we focus on maximizing the life span of the entire network, rather than individual nodes. We believe that it is crucial for many applications that there are no sections of the network that have lost connectivity due to critical nodes running out of power. Our previous model, routing based on energy–temperature transformation, RETT, was first introduced in [1]. Details on the algorithmic aspects of RETT were further published in [2]. In both publications, sensor nodes are assumed to be location-aware;
hence they use their location information to associate themselves with a point in the virtual grid clusters. In this work, sensor nodes are assumed to be location-unaware and organized into disjoint random clusters (i.e. a sensor node belongs to exactly one cluster) in a probabilistic way where an initial set of cluster-head nodes are probabilistically selected. Hence, we generalize the function of RETT to include the clustering process in addition to the routing course in sensor networks. Moreover, we extend the function of RETT-gen to determine the location information for each cluster where sensor nodes are locationunaware.
We analyze the effect of different parameters (cluster radius, network size, cluster-heads selection probability, base station location, base station temperature, heat conductivity coefficient) on the performance of RETT-gen in terms of systems lifetime, average energy consumption, average received data messages and network utility.
In this paper, we present RETT-gen, an energy-equalizing clustering and routing protocol designed for wireless sensor networks. The main goal of RETT-gen is to try to evenly distribute the energy load among all the sensor nodes in the network so that there are no overly-utilized sensor nodes that will run out of energy before the others.
To achieve this goal, RETT-gen uses heat conductivity as a metaphor and uses the heat dissipation difference equations.
In RETT-gen, we transform the expected lifetime of each sensor node to an equivalent temperature, and then by using the heat dissipation equations, we find the hottest path for sending data to the base station, which will not always be the shortest path. Hence, RETT-gen uses both the energy and the distance as routing metrics for routing data packets towards the base station.
The consideration of energy efficiency, lifetime, scalability and adaptability for the RETT-gen protocol makes obligatory the use of a distributed organization. This implies that there should be no centralized unit in charge of most of the tasks for organization of the sensor network.
In a self-organized wireless sensor network, the sensor nodes are not only forwarding packets, they also take part in the network operation. Hence, RETT-gen partitions the network into clusters, each with its own cluster-head node.
Each cluster head is responsible for processing the data, sending the data and participating in routing decisions.
The duty of being a cluster head is shared among all sensor nodes within the cluster. RETT-gen enables each cluster head to aggregate the data coming from the sensor nodes in order to eliminate data redundancy, reduce the amount of data that needs to be sent to the base station and minimize the energy consumption while sending the data.
Moreover, RETT-gen gives the remaining sensor nodes, which we refer to as regular sensor nodes, the opportunity to sleep while not sensing.
2.1. Energy consumption model
Energy efficiency is one of the most important design constraints in wireless sensor network architectures [4].
The lifetime of each sensor nodes depends on its energy dissipation. In applications where the sensor nodes are totally dependent on nonrechargeable batteries, sensor nodes with exhausted batteries will cease operation. A typical sensor node consists mainly of a sensing circuit for signal conditioning and conversion, a digital signal processor, and radio links [5,6]. Hence, during the life cycle of the sensor node, each event or query will be followed by a sensing operation, performing necessary calculations to derive a data packet and send this packet to its destination. Thus, we divide the energy consumption model into the following submodels; the communication energy consumption model, followed by the computation energy consumption model and finally the sensing energy consumption model.
2.1.1. Communication energy consumption
In most applications, sensor nodes are required to communicate to transfer the collected data to one or more base stations. Communication is usually the main source of energy consumption in sensor nodes, which greatly depends on the distance between the source and the destination of the communication link [4]. The radio transceiver typically uses power control in order to expend the minimum required energy to reach the intended recipients. The transceiver can also be turned off to avoid receiving unintended transmissions. In our simulations, we use the radio transceiver model used by [19].
4.1. Simulation experiments setup and goals In this section, we evaluate the performance of RETT-gen protocol via simulations. Unless otherwise specified, we assume 400 sensor nodes are randomly scattered into the field with dimensions 200 200 and a base station located at position x = 210, y = 100. We set the minimum probability (d) for the sensor nodes which are cluster-heads to 4%.
Every result shown is an average of 25 experiments. Each experiment uses a different randomly-generated topology, where each sensor node is assigned a different randomly generated position within the area of interest. In addition, all events are randomly generated.
To assess the performance of RETT-gen, we simulate RETT performance using MATLAB. Throughout the simulations, we consider a static and homogeneous sensor network.
In our simulations, we assume each sensor node is assigned an initial energy of 0.5 J and the data message size for all simulations is fixed at 500 bytes. As a data delivery model, we simulate an event-driven network in which sensor nodes report information only if an event of interest occurs. Furthermore, energy is consumed whenever a sensor node forms or joins a cluster, elects a new cluster-head, senses an event, transmits or receives data, performs data aggregation, and finally when a cluster-head node exchanges messages with sensor nodes in the cluster or with other cluster-head nodes.
DOWNLOAD LINK:
CLICK ME
abstract:
Wireless sensor networks are composed of a large number of sensor nodes with limited energy resources. Once deployed, the sensor nodes are usually inaccessible to the user, and thus replacement of the energy resource is not feasible. Hence, energy efficiency is a key design issue that needs to be enhanced in order to improve the life span of the entire network. Several routing protocols have been proposed to improve the effective lifetime of the network with limited energy supply. In this paper, we propose routing based on energy–temperature transformation, RETT-gen, a scalable energy-efficient clustering and routing protocol designed for wireless sensor networks. The main goal of RETT-gen is to evenly distribute the energy load among all the sensor nodes in the network so that there are no overly-utilized sensor nodes that will run out of energy before the others. To achieve this goal, RETT-gen uses heat conductivity as a metaphor and uses the heat dissipation difference equations. In RETT-gen, we transform the expected lifetime of each sensor node to an equivalent temperature, and then by using the heat dissipation equations, we find the hottest path for sending data to the base station, which will not always be the shortest path. We evaluate the performance of the RETT-gen protocol via simulations, and compare it to the performance of well-known routing protocols (i.e. LEACH [W. Heinzelman, A. Chandrakasan, H. Balakrishnan, Energy-efficient communication protocol for wireless microsensor networks, in: Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS’00), 2000.] and EEUC [C. Li, M. Ye, G. Chen, J. Wu, An energy-efficient unequal clustering mechanism for wireless sensor networks, in: Proceedings of the International IEEE Conference on Mobile Adhoc and Sensor Systems (MASS), 2005.]). Simulation results show that by equalizing the sensor nodes energy, RETT-gen insures that the lifetime of the entire sensor network is maximized, the connectivity in a sensor network is maintained for as long as possible, and that the residual energy of the entire network is of the same order.
Introduction
Wireless sensor networks consist of a large number of densely deployed sensor nodes. The sensor nodes normally feature a low-power wireless radio, an embedded processor, some sensors (i.e., temperature, light, acceleration, etc.) and a small battery. The sensor nodes send their sensed data, usually via radio, to a base station either directly or relayed via other sensor nodes. In order to maximize the lifetime of a sensor network, efficient energy routes need to be selected for the sensor nodes in a reliable manner while operating under strict constraints in computation, communication and energy resources. These unique features have raised some interesting problems that must be addressed when designing a routing protocol for sensor networks. The protocol must be energy efficient, scalable, and adaptable. Moreover, the energy resources of each individual sensor node must be managed effectively by the protocol.
In this work we focus on maximizing the life span of the entire network, rather than individual nodes. We believe that it is crucial for many applications that there are no sections of the network that have lost connectivity due to critical nodes running out of power. Our previous model, routing based on energy–temperature transformation, RETT, was first introduced in [1]. Details on the algorithmic aspects of RETT were further published in [2]. In both publications, sensor nodes are assumed to be location-aware;
hence they use their location information to associate themselves with a point in the virtual grid clusters. In this work, sensor nodes are assumed to be location-unaware and organized into disjoint random clusters (i.e. a sensor node belongs to exactly one cluster) in a probabilistic way where an initial set of cluster-head nodes are probabilistically selected. Hence, we generalize the function of RETT to include the clustering process in addition to the routing course in sensor networks. Moreover, we extend the function of RETT-gen to determine the location information for each cluster where sensor nodes are locationunaware.
We analyze the effect of different parameters (cluster radius, network size, cluster-heads selection probability, base station location, base station temperature, heat conductivity coefficient) on the performance of RETT-gen in terms of systems lifetime, average energy consumption, average received data messages and network utility.
In this paper, we present RETT-gen, an energy-equalizing clustering and routing protocol designed for wireless sensor networks. The main goal of RETT-gen is to try to evenly distribute the energy load among all the sensor nodes in the network so that there are no overly-utilized sensor nodes that will run out of energy before the others.
To achieve this goal, RETT-gen uses heat conductivity as a metaphor and uses the heat dissipation difference equations.
In RETT-gen, we transform the expected lifetime of each sensor node to an equivalent temperature, and then by using the heat dissipation equations, we find the hottest path for sending data to the base station, which will not always be the shortest path. Hence, RETT-gen uses both the energy and the distance as routing metrics for routing data packets towards the base station.
The consideration of energy efficiency, lifetime, scalability and adaptability for the RETT-gen protocol makes obligatory the use of a distributed organization. This implies that there should be no centralized unit in charge of most of the tasks for organization of the sensor network.
In a self-organized wireless sensor network, the sensor nodes are not only forwarding packets, they also take part in the network operation. Hence, RETT-gen partitions the network into clusters, each with its own cluster-head node.
Each cluster head is responsible for processing the data, sending the data and participating in routing decisions.
The duty of being a cluster head is shared among all sensor nodes within the cluster. RETT-gen enables each cluster head to aggregate the data coming from the sensor nodes in order to eliminate data redundancy, reduce the amount of data that needs to be sent to the base station and minimize the energy consumption while sending the data.
Moreover, RETT-gen gives the remaining sensor nodes, which we refer to as regular sensor nodes, the opportunity to sleep while not sensing.
2.1. Energy consumption model
Energy efficiency is one of the most important design constraints in wireless sensor network architectures [4].
The lifetime of each sensor nodes depends on its energy dissipation. In applications where the sensor nodes are totally dependent on nonrechargeable batteries, sensor nodes with exhausted batteries will cease operation. A typical sensor node consists mainly of a sensing circuit for signal conditioning and conversion, a digital signal processor, and radio links [5,6]. Hence, during the life cycle of the sensor node, each event or query will be followed by a sensing operation, performing necessary calculations to derive a data packet and send this packet to its destination. Thus, we divide the energy consumption model into the following submodels; the communication energy consumption model, followed by the computation energy consumption model and finally the sensing energy consumption model.
2.1.1. Communication energy consumption
In most applications, sensor nodes are required to communicate to transfer the collected data to one or more base stations. Communication is usually the main source of energy consumption in sensor nodes, which greatly depends on the distance between the source and the destination of the communication link [4]. The radio transceiver typically uses power control in order to expend the minimum required energy to reach the intended recipients. The transceiver can also be turned off to avoid receiving unintended transmissions. In our simulations, we use the radio transceiver model used by [19].
4.1. Simulation experiments setup and goals In this section, we evaluate the performance of RETT-gen protocol via simulations. Unless otherwise specified, we assume 400 sensor nodes are randomly scattered into the field with dimensions 200 200 and a base station located at position x = 210, y = 100. We set the minimum probability (d) for the sensor nodes which are cluster-heads to 4%.
Every result shown is an average of 25 experiments. Each experiment uses a different randomly-generated topology, where each sensor node is assigned a different randomly generated position within the area of interest. In addition, all events are randomly generated.
To assess the performance of RETT-gen, we simulate RETT performance using MATLAB. Throughout the simulations, we consider a static and homogeneous sensor network.
In our simulations, we assume each sensor node is assigned an initial energy of 0.5 J and the data message size for all simulations is fixed at 500 bytes. As a data delivery model, we simulate an event-driven network in which sensor nodes report information only if an event of interest occurs. Furthermore, energy is consumed whenever a sensor node forms or joins a cluster, elects a new cluster-head, senses an event, transmits or receives data, performs data aggregation, and finally when a cluster-head node exchanges messages with sensor nodes in the cluster or with other cluster-head nodes.
DOWNLOAD LINK:
CLICK ME
No comments:
Post a Comment