On the interest of artificial intelligence approaches in solving the IoT coverage problem

This survey deals with the 3D indoor deployment in IoT collection networks to identify the right locations of the IoT connected objects and, subsequently, to manage the coverage holes while guaranteeing other objectives such as localization and connectivity of IoT devices. These coverage holes result generally from the initial random distribution of the objects. Unlike the existing studies, the present survey does not only focus on the use cases and applications of the different deployment issues but, it also analyzes the problem of deployment by showing its different models and hypotheses presented in the literature and highlighting the used evaluation and performance criteria. The aim of the study is to highlight the relevance of artificial intelligence techniques; especially meta-heuristics, evolutionary optimization, machine learning and reinforcement learning; in finding coverage solutions, better than other methods. Through experimental validation, the performances of the various deployment and redeployment approaches in a 3D indoor context are compared. These approaches involve computational geometry, virtual force, clustering, mathematical modeling, and evolutionary optimization-based approaches. Afterwards, they are investigated and categorized according to their specifications into different summary tables. Statistical and complexity tests are performed to evaluate the complexity of these approaches. In addition, current 3D deployment trends are presented, and some outstanding deployment issues are discussed. Based on the experimental and simulation findings, the behaviors of evolutionary optimization algorithms are compared to those of the other deployment techniques. The obtained findings demonstrate that using artificial intelligence, specifically many-objective optimization algorithms for 3D deployment of IoT networks is more beneficial as it allows enhancing the quality of deployment compared to the other deployment approaches.

Automated summary

This survey focuses on the 3D indoor deployment in IoT collection networks, specifically on finding the right locations of IoT connected objects and addressing coverage holes. It analyzes different models and hypotheses presented in the literature, and highlights the use of artificial intelligence techniques such as meta-heuristics, evolutionary optimization, machine learning, and reinforcement learning. The study demonstrates that using these techniques for 3D deployment of IoT networks can enhance the quality of deployment compared to other approaches.