Subject Training

The Internet of Things (IoT) now plays a significant role in our daily lives. Various sensors and other connected devices are used to measure and transmit data (temperature, humidity, etc.) to users, who then use this data for decision-making or monitoring. To ensure optimal data transmission performance (loss rate, latency, throughput, etc.), several communication technologies can be used (Wi-Fi, Zigbee, LoRa, etc.). Among these, the IEEE 802.15.4 standard is frequently used in IoT networks due to its advantages in terms of low power consumption and sensor network management. This standard relies on the CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) protocol to regulate access to the radio channel. This protocol works by avoiding collisions through prior channel listening before any transmission and the introduction of random delays when the channel is busy.

In the process of evaluating the performance of a communication technology before its deployment, simulation is commonly used to reduce costs and obtain reliable results. However, it is essential to compare the results obtained in simulation (using tools like Cooja [1]) with those obtained in real-world experimental environments such as FIT IoT-Lab [2]. This allows for a better understanding of the differences between simulated models and real-world environments, particularly for key parameters such as reliability and latency. In this context, the concept of a "Digital Twin (DT)" aims to faithfully reproduce the behavior of a physical system using data-oriented modeling. Coupling experimentation and simulation makes it possible to obtain an accurate digital twin of the network [3].

This internship aims to study the differences and similarities between a simulation environment (Cooja) and an experimental environment (FIT IoT-Lab) regarding the performance of the CSMA/CA protocol used in the IEEE 802.15.4 standard, in order to design a digital twin of a network. The focus will be on the impact of configuration parameters (backoff, contention window, etc.) on network performance, particularly reliability and latency. The identified tasks are as follows:

⬔ Familiarization with the Cooja simulator and the FIT IoT-Lab platform, ⬔ Use of CSMA/CA protocol implementations in both environments, ⬔ Comparative analysis of network performance in terms of reliability and latency, ⬔ Integration of artificial intelligence methods to: (i) Automatically calibrate simulation models (e.g., by adjusting radio propagation parameters), (ii) Predict real-world performance from simulations using supervised learning techniques.

References: [1] Osterlind, F., Dunkels, A., Eriksson, J., Finne, N., & Voigt, T. (2006, November). Cross-level sensor network simulation with cooja. In Proceedings. 2006 31st IEEE conference on local computer networks (pp. 641-648). IEEE. [2] Adjih, C., Baccelli, E., Fleury, E., Harter, G., Mitton, N., Noel, T., ... & Watteyne, T. (2015, December). FIT IoT-LAB: A large scale open experimental IoT testbed. In 2015 IEEE 2nd World Forum on Internet of Things (WF-IoT) (pp. 459-464). IEEE. [3] Si-Mohammed, S., Bardou, A., Begin, T., Lassous, I. G., & Vicat-Blanc, P. (2024). Ns+ ndt: Smart integration of network simulation in network digital twin, application to iot networks. Future Generation Computer Systems, 157, 124-144.