WirelessHART is an emerging wireless sensor network protocol. In this paper, a joint graph routing algorithm for maximizing the network lifetime (JRMNL) in WirelessHART is proposed. Node communication load factor is approximately estimated by matrix operations for the first time. Then node communication load, the residual energy, and the link transmission power are integrated as a link cost function that is accurately measured in this algorithm. A node chooses the optimal next hop by comparing the link cost function of all its neighbor nodes, which guarantees the energy balancing of the whole network. Simulation results show that the proposed algorithm can extend network lifetime by a factor of 2 relative to the maximum residual energy selection algorithm and prolong the network lifetime by a factor of 7 relative to the minimum transmission power routing algorithm, but the average energy consumption per route will increase by 2 dBm compared with the minimum transmission power routing algorithm.
WirelessHART is the first international protocol aiming for industrial automation and process control. Since it was announced in 2007, many scholars have carried out deep exploration into the application, the key technologies, and the protocol stack design of WirelessHART [
Different from the star network or the tree network in which data focuses on some fixed nodes, the topology of WirelessHART is a mesh network, in which data can be transmitted on different paths. In WirelessHART network, every sensor node is a router, and WirelessHART defines two kinds of routing named graph routing and source routing, respectively. The source routing is used for path detection, while the graph routing can provide abundant redundant path to improve the reliability of the wireless communication.
It is tested that most of a node’s energy is used for wireless data transmission, so how to choose the appropriate route to make the maximum numbers of transmission when uploading data to the gateway is a meaningful topic. In recent years, [
Based on the graph routing network topology in WirelessHART, this paper proposes a Joint Routing Algorithm for Maximizing Network Lifetime (JRMNL) by taking into account the remaining energy and communication load of neighbor nodes as well as the transmission power of the route. This paper has proven that JRMNL can effectively reduce the energy consumption of a single route and maximize the network lifetime at the same time.
The rest of this paper is organized as follows. Section
WirelessHART network is composed of a lot of stationary and randomly distributed sensor nodes in a certain range, and the gateway serves as the sole root node for data aggregation and divergence. In order to achieve the high reliability in data transmission, this paper adopts the graph routing, by which each node has two or more neighbors so that redundant paths are provided to improve the reliability of communication.
In graph routing, the next-hop destination address is referred to as the father nodes or the neighbor nodes who serve as the relays, and data should be delivered to the gateway through a lot of relays. The establishment of the network topology is shown in Figure
Graph routing topology structure.
In this paper, the quasistatic flat fading channel is considered for wireless communication between nodes, so the channel parameters remain unchanged over a given super frame transmission. Nevertheless, over the different super frame transmissions, channel parameters are variable and are independent complex Gaussian random variables with zero mean and unit variance. In addition, the noise terms are also set to be complex zero mean Gaussian random variables with the variance
Assume a WirelessHART network with
WirelessHART is considered to be the most promising industrial wireless sensor network protocol due to its safety and easier controlling. According to the information such as the location of nodes, the network manager can easily calculate the energy consumption for transmission between hops through the software tools, and then guides the WirelessHART system to find the optimal route based on it.
Based on the proposed graph routing topology structure, we consider several kinds of existing routing algorithms. Assume that node
In addition, the ELHFR algorithm proposed in [
The link connectivity is forwarded to network manager regularly, and the link cost function is computed periodically according to the previous formulas. The calculation results will be broadcasted to all nodes in network updating. The focus of this paper is to find the best link cost function for choosing the next-hop node under the graph routing topology. So this strategy must make the maximum numbers of wireless data transmission, namely, the maximum network lifetime. In our study, the best link cost function can and only can be achieved when some important elements that will be introduced in the next part are specially considered.
Since WirelessHART network adopts the hierarchical topology, the number of hops for a node to reach the gateway is the degree of the node, which is a fixed number. It is on this premise that studying the path selection criterion in each jump becomes very meaningful. Without considering the energy consumption for data processing and information broadcasting, the network lifetime can be expressed as the number of data forwarding to the gateway. For the path selection criterion, the residual energy and the energy consumption for transmission are clearly two determinants, but it is not enough. In the actual WirelessHART network, according to formula (
WirelessHART network is a changing one, the harsh environment will lead to node failure at any time, and the node’s remaining energy also changes all the way. Therefore, the best routing should also be changed all the way. In order to solve the load balancing problem in the changing network, we introduce the concept of the node communication load. Communication load refers to the communication frequency of nodes in a certain period of time, and it indicates the node’s intensity of wireless communication. Obviously, nodes with the lower communication load should be preferentially selected. Therefore, we need to analyze the residual energy, the transmission power, and the node communication load and then enable them to be nonlinear combined into a composite cost function. The power exponents of three factors as well as their respective representative algorithms are as shown in Table
Determinant features.
Determinant | Power exponent | Representative algorithm |
---|---|---|
Node communication load |
|
None currently |
Node residual energy |
|
ELHFR |
Link transmission energy consumption |
|
MPCR |
The joint routing algorithm for maximizing network lifetime JRMNL coalesces the node communication load, the residual energy, and the link transmission power together as a whole and chooses the optimal next hop by comparing the link cost function. The formation of the link cost function is divided into two steps, which are the calculation of node communication load factor and the nonlinear combination of the factor, the residual energy and the transmission power consumption.
As the name implies, the node communication load factor reflects the communication frequency of nodes. The greater the load factor is, the higher the communication frequency is and the faster the energy consumption is. When the graph routing topology structure is established, the network manager will soon label all the nodes and links. The calculation process for load factor is as follows.
Without considering retransmission and the energy consumption for data processing, we use MATLAB software to simulate JRMNL, ELHFR, and MPCR under the same conditions. According to the system and channel model built in Section
Simulation parameters.
Parameter | Quantity |
---|---|
Number of nodes |
|
Square area m2 |
|
Gateway coordinate |
|
Noise variance |
−70 dBm |
|
Unsure |
Path loss exponent |
2 |
Initial energy |
1 |
|
0 |
Transmission rate |
2 b/s/Hz |
Throughput |
1.9 b/s/Hz |
According to Table
To simplify the experiment, we can fix the value of
Determinant definitional domain.
Power exponent | Value range |
---|---|
|
|
|
|
|
|
After building the layered graph routing topology, we randomly select the source nodes that need to find a path to the gateway and then use three kinds of routing algorithms for routing until all nodes in the network are running out of energy. In the actual WirelessHART network, due to the limited bandwidth resource and the real-time demand of data transmission, the number of sensor nodes is limited. Generally speaking, in the light of the perishing industrial application environment, the network permits 50 to 100 devices. Therefore, in our experiment, we choose up to 100 nodes [
Network lifetime versus the number of nodes based on ELHFR and MPCR.
As can be seen from the figure, when the number of nodes is larger than 40, MPCR algorithm can extend the network lifetime by a factor of 2 relative to ELHFR algorithm. It shows that the energy consumption for transmission plays a more important role than the residual energy in the routing selection process. Therefore, in the link cost function, the value of
Network lifetime versus the number of nodes based on JRMNL when
We can see from the figure that the difference between their network lifetimes is very small. When the number of nodes is larger than 50, the network lifetime when in case
In the above experiments, a measure of the cost function in JRMNL is verified by the experiments on the accuracy of the values of the power exponents, and it is verified that JRMNL can maximize the network lifetime. Now, we will further study the advantages of JRMNL compared with other routing algorithms and the optimization between the network lifetime and the power consumption.
Finally, under the same simulation environment, we get the average energy consumption per route and the average network lifetime based on different
Network lifetime versus the number of nodes based on three routing algorithms.
Average transmission power per route versus the number of nodes based on three routing algorithms.
Network lifetime versus different desired throughputs based on three routing algorithms when
As can be seen from Figure
Figure
Figure
Aiming at the highly reliable WirelessHART protocol, this paper proposes a joint routing algorithm for maximizing network lifetime JRMNL. Based on the graph routing topology structure with high reliability, the algorithm builds a unique link cost function by a nonlinear combination of several parameters including the transmission power, communication load factor, and the residual energy of nodes. A node will select the optimal next hop by comparing the link cost function of all the possible links. JRMNL can greatly increase the network lifetime by 7 times and 2 times compared with ELHFR and MPCR, respectively. When the throughput is small the algorithm can also have a relatively longer network lifetime. In addition, since JRMNL takes into account a variety of important determiners, the average energy consumption per route is about 2 dBm larger than MPCR but is nearly 4 dBm smaller than ELHFR. Due to the particularity of the WirelessHART protocol, the combination of the optimal routing and the link scheduling will be an interesting issue and is worthy of further study.
This research was supported by the National Science and Technology Major Project of the Ministry of Science and Technology of China (2011ZX03004-001-01) and was under the Modern Communication Team support program supervised by Shenzhen Key Laboratory for Information Science and Technology in Tsinghua University.