*1.3. Significance for Study of Energy in Wireless Sensor Networks*

To evaluate the network performance, one may consider parameters that evidence proper network operation directly influencing the energy consumption of each node. There are local and global parameters. Global parameters display the total energy costs for the network considering each type of energy for each specific activity. In contrast, local parameters provide total energy consumption rates for a single node. This energy depends on the location of the node within the topology regardless of how near or far they are located from the coordinator node and how much traffic is transmitted through it [31].

An energy-efficient routing protocol decreases the consumption of the nodes by routing data through paths that display the least amount of energy. There are some special mechanisms to achieve this goal such as optimization of jumps to the destination node, maintenance of optimal and valid routes, reduction of transmission delays, and reduction of packet retransmissions and attempts to listen to the channel [32,33]. In this paper, we assess the aforementioned aspects to compare against four sensor network protocols.

Concerning the communication channel, it is a factor that significantly influences the energy consumption because the protocol executes a series of listening attempts to determine whether the channel is already busy with other information packets. The carrier senses multiple accesses with the collision avoidance (CSMA/CA) protocol [34] and works as follows: first, a node begins listening to the wireless channel and if it is free, the node begins transmitting. If the wireless channel is not free, the node recalculates a random delay, waits, and listens again. We are using the MAC-level protocol used for all extensions of 802.15.4 (including the original version), which is the CSMA/CA that guarantees a high data rate. A network recognition is being carried out at all times to check the status of the channel (carrier detection). Only when free, data can be sent. In the 802.11 standard, the physical layer polls the energy level over the radio frequency to determine whether or not there is transmission. If the channel is busy, a random timer starts (with a maximum of five back off periods), the timer only discovers time with free channel, transmits when it expires, and finally, if it does not receive ACK, it increases the back off. This metric is known as CSMA/CA retries. If these CSMA/CA retries are frequent, the channel is busy most of the time. Consequently, there might be several collisions due to overload. In addition, when the wireless channel is permanently busy with information packets, there are many collisions and retransmissions of packets. This fact influences energy consumption because the nodes spend more time and capacity retransmitting over and over.

In a network layer, overloads are an important factor that influence energy consumption. The efficiency of the routing protocol may also be measured by the number of packets the protocol needs to route to its destination. A protocol with many control packets will contribute to packet collisions and overall performance reduction. In terms of route discovery, in all the protocols considered, the nodes exhibit capacity to know their neighbors. In AODV and DSR, the nodes update the routes as required. However, in ZTR and MPH routing, nodes periodically update routes. Known neighbors helps nodes to establish valid destination routes to forward messages and to reduce the number of retransmission packets required. The metric named 'valid routes' consists of paths that are not damaged and these routes are immediately available and ready to be used for sending information. If nodes are able to handle several valid routes, it is more likely that packets will be forwarded and will not be retransmitted continuously, thus saving energy resources.

Another important metric is found in the tables where the routes to different network destinations are located. These routes may either be valid or obsolete. Hence, their maintenance is essential to prevent loss of information within the network. Network energy consumption is directly related to the complexity in the administration of routing or neighbor tables. As sensors execute huge routing processes, energy consumption increases if these routes have not been properly updated. This is why it is also important to assess route delays; they are directly related to the number of jumps that a node takes to reach a destination.

This work seeks to further advance the study of performance metrics in the analysis of sensor networks. One of the most important contributions of this work is the relationship between metrics and their influence on energy consumption in a WSN, as well as the comparison of how these metrics perform in widely known routing protocols with respect to a protocol proposed by the authors. An important assessment of these parameters is that routing tables provide critical information about network reliability and the number of valid routes available to the nodes for sending information. These metrics will directly influence packet loss rates in the network as well as information delivery reliability. The aforementioned study is implemented in a sensor network for each routing protocol in which the proposed energy model is also implemented, as described in Table 1. This model is based

on the operation of the Texas Instruments CC2530 sensor [35] in active mode and stable conditions. Besides, for nodes under LoRa technology and WiFi, we use the model described in [36].


**Table 1.** Energy model base for each of the main tasks in a node [37].

Global parameters denote the energy rates that each node spends for all activities performed within the network. These types of energy include the energy consumed by the microcontroller, which is regulated by the time the sensor remains on; the CSMA energy used when the node is listening to the channel to determine whether packets may be transmitted or received and the node is executing the CSMA/CA algorithm; the switching energy consumed when switching between activities, i.e., changing from transmission to reception or vice versa; transmission and reception energy used to transmit or receive a packet, respectively; awakening energy used when the node turns on; and the shutdown energy consumed when the node turns off.

All of these energies will promptly report how nodes save or consume energy, as the case may be, together with the routing protocol. Each type of energy is provided as an overall network metric. In other words, each energy is incorporated to observe possible points of interest or zoning in the network.

Local energy balance depends on the proximity of a node to the destination node. In the scope of this work, the final destination is the coordinator or collector node. As a local aspect, the energy consumption at each node is calculated, which denotes consumption depending on the distance between the node and the coordinator. Thus, the nodes closest to the collector will send their packets and also retransmit packets coming from other nodes. For this reason, the relational position of a node within the network significantly matters.
