Public transport systems often face significant revenue losses due to fare evasion, affecting operational efficiency and service quality. Traditional methods of detecting fare evasion, such as manual inspections, are often ineffective because of the high volume of passengers and limited staff availability. This study presents a new approach to fare evasion detection by combining behavioural AI, reinforcement learning, IoT sensors, and privacy-conscious technologies. The system incorporates multi-zone ticket validation, AI-powered cameras, and features in a mobile app to monitor passenger behaviour in real-time, ensuring continuous ticket compliance without compromising privacy.Key components of the system include motion sensors, pressure sensors, NFC readers, and a federated learning framework, which help create a seamless and accurate detection system. This system is expected to reduce fare evasion by 15-20%, recovering millions of pounds in lost revenue annually. Additionally, the system will enhance the passenger experience by making ticket validation easier and reducing congestion. Overall, this solution offers a scalable, efficient, and ethical way to improve the performance and sustainability of public transport systems.