Machine Learning for Multi-access Edge Computing and Autonomous Driving

Last twenty years have seen the explosive growth of information technology, and we have stepped into the era of information explosion. These technologies significantly contribute to the generation of the unprecedented quantity of data and make it challenging for data storage and data processing. Multi-access edge computing is emerged as an effective way to relieve the pressure of data processing. One of the typical edge computing assisted applications is autonomous driving, which relies heavily on edge servers for data analysis and decision making. However, with the growth of end users, how to allocate the available edge computing resources to fulfill the requirements becomes a challenging problem. Machine learning, as a popular and effective method for decision making, has diverse applications. Machine learning is the study of computer algorithms that can improve automatically through experience and by the use of data. Machine learning algorithms build a model based on sample data, known as training data, in order to make predictions or decisions without human interference. Therefore, there is a great potential to utilize the ideas, methods, and models of machine learning to make decisions for edge computing resource allocation. Given this background, this dissertation provides a theoretical research between machine learning, multi-access edge computing, and autonomous vehicles. Especially, different machine learning models and edge computing assisted applications are discussed. The main contributions of this dissertation are as follows. 1. The basic concepts and classifications of machine learning are provided. The architecture, characteristics, and key technologies of multi-access edge computing are given as well. 2. Applications of machine learning for edge computing resources allocation are studied. Also, except the utilization of machine learning method, the efficient machine learning framework design is also discussed. 3. Numerical results are provided to show that the proposed method can be utilized for object realization in edge computing scenario such as accurate prediction, low latency, etc. 4. The potentials of machine learning for multi-access edge computing applications in future wireless networks are discussed. This dissertation provides a theoretical research between machine learning, multi-access edge computing, and autonomous vehicles, in which different machine learning models are utilized in different multi-access edge computing assisted applications as well as efficient machine learning mechanism design. This work places a fundamental research on edge computing resource allocation. This research has the potential to contribute to the future wireless network area and has a long term effect on problems such as edge computing resource deployment, service guaranteed wireless network design.