Integration of blockchain and machine learning in autonomous driving

In recent years, the transportation industry has witnessed remarkable advancements in autonomous driving systems. The fusion of blockchain and machine learning technologies can improve the way for the development of safer, more reliable, and more efficient autonomous vehicles. Blockchain can be employed to store, manage, and share the amounts of data produced by the sensors and cameras of autonomous cars, ensuring data integrity and security. Simultaneously, machine learning algorithms can analyze this data in real time, enabling the vehicle to make well-informed decisions for navigation and adaptation to changing circumstances. This combination of technologies holds the potential to enhance safety, performance, and scalability in autonomous car systems, making them a more viable and appealing option for consumers and industry stakeholders. This thesis explores the integration of machine learning and blockchain technologies in autonomous vehicles. The thesis proposes a novel approach for connecting autonomous vehicles through blockchain, allowing them to exchange obtained learning experiences during the training process, which utilizes Neuro Evolution of Augmenting Topologies (NEAT) algorithm with reinforcement learning. The results show that the proposed approach has valuable insights into the future of autonomous vehicle training and offers a more efficient and effective method for storing this data through blockchain technology making the training process of autonomous vehicles much faster and more efficient according to many different evaluation metrics.