Humans can be recognized by their distinctive walking patterns, which have been established using a variety of techniques, including the use of sensors. Footstep recognition, which analyzes the distinctive characteristics of a person's footsteps, can be applied in a range of scenarios, including security, criminal investigations, human behavior security applications, and healthcare for monitoring and analyzing gait abnormalities. This paper discusses the most recent work on footstep analysis and identification systems in terms of using the floor-based approach. It explains the various artificial intelligence methods as well as the machine learning and deep learning algorithms applied to the recognition and analysis of footsteps, the various feature engineering techniques applied to each type of sensor, the affection of the engineered features on the footstep identification and analysis systems, and the best suitable features for each type of sensor and application, which provide researchers in this domain with an appropriate grounding in footstep identification and analysis utilizing the floor-based technique.