Energy efficiency is a paramount concern in IoT networks, directly affecting the network's scalability, reliability, and sustainability. Several factors contribute to energy consumption in IoT devices, including:

I. Radio transmission: the transmission of data over wireless channels consumes a significant amount of energy.

II. Sensing and processing: collecting and processing sensor data requires energy for both the sensing hardware and the processing unit.

III. Idle state power consumption: even when idle, IoT devices consume a certain amount of energy to maintain their functionality [2].

Energy constraints Limited battery capacity of IoT devices Scalability Handling a large number of devices and data traffic Security and privacy Protecting sensitive data from unauthorized access Reliability and fault tolerance Ensuring continuous operation in the presence of failures

Efficient energy management strategies are essential to address these challenges. One crucial aspect of energy-efficient IoT networks is the design of robust and energy-aware routing protocols. Routing protocols determine the optimal paths for data transmission, significantly impacting the overall energy consumption of the network.

Low-Energy Adaptive Clustering Hierarchy (LEACH), one of the pioneering energy-efficient protocols for WSNs, uses clustering to reduce energy consumption by rotating cluster heads (CHs) to distribute the energy load evenly. LEACH operates in rounds: during each round, sensor nodes elect CHs that aggregate and transmit data to the base station. The clustering and data aggregation techniques significantly reduce the frequency of direct communications to the base station, which conserves energy.

I. Strengths: LEACH improves network lifetime and balances energy consumption by clustering nodes. II. Limitations: its random CH selection may lead to uneven distribution of CHs, and its performance is limited in larger networks due to scalability issues [4].

Power-Efficient Gathering in Sensor Information System (PEGASIS) improves upon LEACH by creating a chain of nodes. Each node transmits data to its closest neighbor, and one node in the chain sends the aggregated data to the base station. This technique reduces the number of direct communications with the base station, thereby conserving energy. PEGASIS further reduces the number of transmissions by allowing only one node per round to communicate directly with the base station.

I. Strengths: high energy efficiency due to minimal transmission distances and data aggregation.

II. Limitations: the protocol introduces delays due to sequential transmissions along the chain, making it less suitable for time-sensitive applications in IoT networks.

Threshold-sensitive Energy Efficient Network (TEEN) is designed for time-critical applications, where energy efficiency is achieved through a threshold-based data transmission strategy. Nodes transmit data only when certain threshold conditions (such as temperature or humidity) are met. This method reduces unnecessary transmissions, thus conserving energy, which is essential in event-driven smart city IoT applications, such as emergency response systems.

I. Strengths: high energy efficiency in threshold-based, event-driven applications; reduces unnecessary transmissions.

II. Limitations: ineffective for applications requiring periodic updates, limiting its use in real-time smart city scenarios where constant monitoring is essential.

Adaptive Threshold-sensitive Energy Efficient Network (APTEEN) extends TEEN by providing periodic and threshold-based data reporting, making it suitable for hybrid smart city IoT applications requiring periodic updates and event-based reporting. APTEEN addresses some of its limitations by adapting to a broader range of applications, though it introduces additional overhead in managing hybrid thresholds.

I. Strengths: flexibility to support periodic and event-driven data, improving adaptability for smart city applications.

II. Limitations: increased complexity and overhead from managing adaptive thresholds and hybrid reporting methods [5].

Energy-aware Routing Protocol (EAR) is a protocol that focuses on routing based on the remaining energy of the nodes. Nodes with higher energy levels are preferred in routing decisions, prolonging the network's lifetime. This approach benefits IoT networks in smart cities where certain nodes may have more energy due to placement or usage patterns.

I. Strengths: increases network longevity by prioritizing high-energy nodes.

II. Limitations: this may lead to uneven energy consumption and potential isolation of low-energy nodes, especially in dynamic IoT networks [6].

This algorithm organizes nodes into clusters to reduce the number of direct transmissions to the base station, thereby conserving energy.

I. Input: set of sensor nodes N, maximum cluster radius R II. Output: clusters formed with designated CHs for each node i in N:

if i is not part of any cluster:

Create a new cluster C Add node i to cluster C For each node j in N:

if distance(i, j) ≤ R and j is not part of any cluster:

Add node j to cluster C Select the node with the highest residual energy in C as the CH end for This algorithm ensures that nodes within a specified radius are grouped into clusters, with one node in each cluster chosen as the CH [11].

CH selection is based on residual energy to distribute energy consumption across nodes and extend network lifetime evenly.

I. Input: cluster C, residual energy of each node E(i) in C II. Output: selected CH Initialize CH = nullInitialize max_energy = 0 for each node i in C:

if E(i) > max_energy:

max_energy = E(i) CH = iend for III. Result: node CH is designated as the CH based on the highest residual energy, ensuring balanced energy depletion across the network [12].

This algorithm reduces energy consumption by transmitting data only when it meets a predefined threshold, conserving power for non-critical data.

I. Input: sensed data D, predefined threshold T II. Output: aggregated data sent to the base station if the threshold is met for each node i in cluster C:

if |D(i) - previous_D(i)| > T:

Aggregate data in C Send aggregated data from the Cluster Head to the base station else:

Skip transmissionend for This approach helps reduce unnecessary data transmissions, significantly conserving energy in event-driven smart city applications [9].

A machine learning model predicts the optimal route by analyzing historical network data, including energy levels, network congestion, and latency, to find the most energy-efficient path for data transmission.

I. Input: network data D (e.g., energy levels, distance, congestion levels) II. Output: predicted optimal route R for data transmission Train ML model (e.g., Reinforcement Learning, Decision Tree) on historical network data D for each transmission request:

Input current state parameters (energy, distance to base, congestion) into the ML model Predict optimal route R based on model output Send data along route Rend for I. The machine learning model is pre-trained offline on historical data and periodically updated with new data to ensure adaptability to changing network conditions.

II. The model outputs a recommended route that optimally balances energy consumption and transmission efficiency [13].

To prevent the rapid depletion of certain nodes, this algorithm redistributes tasks among nodes with higher energy, reconfiguring clusters as necessary.

I. Input: CH, residual energy of nodes E(i) II. Output: updated clusters with re-assigned roles for each CH:

if E(CH) < threshold:

Select a new CH with higher residual energy in the cluster Reconfigure the cluster to balance energy load end if end for This algorithm dynamically updates CHs and roles within the network to extend the network lifetime and maintain balanced energy consumption [14].

I. Observation: the proposed protocol reduced overall energy consumption compared to the other protocols, primarily due to clustering, threshold-based data transmission, and machine learning-based routing.

II. Result: energy consumption was reduced by approximately 30% compared to LEACH and by 20% compared to PEGASIS and TEEN.

I. Observation: the proposed protocol extended network lifetime significantly due to balanced energy consumption across nodes.

II. Result: network lifetime increased by about 40% compared to LEACH and 25% compared to PEGASIS and TEEN.

I. Observation: the packet delivery ratio remained consistently high, outperforming the other protocols. II. Result: the proposed protocol maintained a packet delivery ratio of about 95%, while LEACH, PEGASIS, and TEEN averaged around 88–90%.

I. Observation: the protocol's latency was comparable to PEGASIS and TEEN and slightly higher than LEACH due to threshold-based data transmission.

II. Result: latency remained acceptable for smart city applications and was suitable for periodic and event-based data.