With the increasing importance of data rights and privacy preservation, federated learning, as a new machine learning paradigm, can achieve the goal of solving data silos as well as privacy preservation problems without exposing the data of all parties. This paper provides an exhaustive and systematic review of federated learning, highlighting its concepts, aggregation methods, applications, and challenges. First, we introduce the basic concepts of federated learning, including the principles behind it and the basic workflow. Then, we delve into commonly used aggregation methods in federated learning, including federated averaging and optimisation algorithms in federated learning. Next, we discuss in detail the applications of federated learning in various domains, covering a wide range of aspects such as healthcare, finance, and the Internet of Things. Finally, we analyze the challenges facing federated learning, including aspects of privacy protection, communication efficiency, and data heterogeneity. Through this review, we hope to provide a comprehensive understanding of the federated learning environment and lay the foundation for subsequent research.