In this project, the main focus is on understanding and detecting Man-in-the-Middle (MITM) attacks, where an unauthorized party intercepts communication between two parties. Facilitated by the usage of a tool called Ettercap to simulate various MITM attacks and carefully analyzed the network traffic using Wireshark, distinguishing between normal and malicious activities, a detailed dataset was created.
The main role involved implementing a type of machine learning called Q-learning within a framework called the Markov Decision Process. This allowed a reinforcement learning (RL) model to learn and make decisions on the best actions to detect MITM attacks. The model was continuously improved to adapt to new types of attacks, enhancing our ability to defend digital systems against evolving MITM threats. Essentially, the usage of RL strengthens the security of digital infrastructure against sophisticated cyberattacks.