Came across a paper that could have very well been generalized (i.e, industry agnostic). Once you start going through the paper, you soon realize that it does not matter if the network is healthcare-focused. The approach suggested will work across industries with a similar IoT network setup. https://lnkd.in/gA_B4Etf

Authors highlight that IoT-based healthcare systems face challenges like high energy consumption, network scalability, latency, and efficient routing. Traditional clustering and routing protocols may not adapt well to dynamic conditions or optimize multiple metrics simultaneously.

These challenges are experienced across most IoT networks, irrespective of industry.

Another rationale provided is that in healthcare settings, reliable and timely delivery of data is critical, so energy efficiency and low latency are especially important. In fact, setting up an extensive IoT network is done primarily for timely capture and reliable delivery of data, no matter what the industry.

But when it comes to the methodology, it seems robust and hence my postulation that it can be considered irrespective of industry.

The authors propose a two-stage method combining Adaptive Fuzzy Logic (AFL) for cluster head (CH) selection and a hybrid Particle Swarm Optimization + Genetic Algorithm (PSO-GA) approach for routing.

A. Adaptive Fuzzy Logic for CH Selection

The algorithm considers four input parameters for each sensor node (SN):
1. Residual energy
2. SN density (how many neighbors)
3. Distance to the base station (BS)
4. Link stability (i.e. reliability of the communication link)

The fuzzy logic membership functions are dynamically adapted over time (each clustering round) based on current network statistics, rather than being fixed a priori. This allows the system to better reflect the changing state of the network (e.g. energy levels changing, topology changes)

A fuzzy rule base maps combinations of these inputs to a CH suitability score, and sensor nodes with the highest scores are elected as CHs in their local vicinity.

B. Hybrid PSO-GA Routing

After CHs are selected, the challenge is to route data from CHs to the BS in a way that balances energy usage, delay, reliability, and throughput.

Each possible route is encoded as a “particle” in PSO. The fitness function combines several factors (energy consumed, delay, link stability, packet delivery ratio) via weights.

To avoid the usual limitation of PSO (getting stuck in local optima), GA operations (selection, crossover, mutation) are periodically invoked to introduce diversity and exploration into the solution pool.

The hybrid method thus leverages PSO’s fast convergence and GA’s global search capability to find better routing paths.