Smart mobility and infrastructure The role ofIntelligent Transportation Systems onaccessibility with regards to advancements in AI

Sammanfattning

The thesis discusses how Artificial Intelligence (AI) may interplay with Intelligent Transportation Systems (ITS) to enhance the four main transport modes of accessibility to pedestrians, cyclists, car users, and public transport users. As the population becomes urbanized and the demand for mobility has grown, the process of developing responsive infrastructure becomes a critical area of study. The research employs a qualitative methodology based on a literature study. Its aim is to identify the limitations regarding ITS today and propose AI-driven innovations. For pedestrians, the thesis discusses intelligent crosswalks linked to adaptive traffic signals. Though intelligent, these systems are limited by their reliance on cameras and LiDAR sensors, which are sensitive to adverse climatic conditions. A solution to this is to have AI systems that are capable of handling multimodal platform-style sensing technology that uses radar, audio, and thermal sensors to sustain accurate detection. Such systems can adapt to the usage of sensors depending on weather conditions to preserve the level of accuracy. For cyclists, ITS tends to give priority to vehicles over bicycle traffic. The thesis suggests that AI can help to achieve predictive systems that observe the demand of bicycle users to give priority to cyclists at intersections. With a federated learning process that empowers cities to train AI systems based on traffic patterns, cyclists would have adaptive solutions based on real-world data from various cities. For automobile traffic, V2X technology from various automobile makers creates a hurdle in successful communication as well as information exchange between the cars and the environment. The Thesis envisions a future where an AI-controlled V2X platform has the potential to standardize communication for every vehicle. Of particular significance in light of self-driving vehicles, which are programmed to function at their maximum capacity with the help of successful information exchange. For the public transport sector, existing ITS solutions are normally operated from centralized control centers, hence less responsive to unexpected disturbances. A decentralized AI system where vehicles communicate and respond to each other independently in real-time can improve the response of the entire system.