The problem IoTVerse solves

In the ever-expanding realm of the Internet of Things (IoT), our focus on the Detection of Botnet Attacks emerges as a critical solution to safeguard interconnected devices. IoT devices, characterized by their vast numbers and connectivity through wireless technologies, face inherent security challenges. Conventional botnet detection, reliant on complex algorithms and M2 chips absent in IoT devices, falls short in addressing the rising threat severity. Our innovative approach introduces pre-cooked Machine Learning models, strategically designed for the unique constraints of IoT environments. While demanding in training computation, these models offer a distinct advantage – minimal prediction computation, optimizing performance for resource-constrained IoT devices. By bolstering security measures, our solution addresses privacy concerns, mitigates service disruptions, and fortifies the infrastructure integrity of IoT ecosystems, ensuring their resilience against the evolving landscape of cyber threats.

Challenges we ran into

Building this project presented several challenges, each demanding unique solutions. The primary hurdle was identifying an algorithm superior to existing options for the Detection of Botnet Attacks on IoT infrastructure. Navigating through the multitude of algorithms available required extensive research and testing to pinpoint the most effective one.

Another significant challenge involved deciding on multiple models and determining which algorithm to deploy. The process involved evaluating various machine learning models tailored for IoT environments. Selecting the most suitable combination required a delicate balance between accuracy and computational efficiency.

Training the models posed another difficulty. The intricacies of handling large datasets, model convergence, and optimizing for IoT's limited computing power demanded meticulous attention. Overcoming these challenges involved iterative testing, collaborative problem-solving, and leveraging expertise in machine learning techniques. Ultimately, the project's success hinged on a combination of research, experimentation, and a strategic approach to algorithm selection and model training.