Abstract. This article provides a review of the literature and existing research in recent years on the topic, describes the tasks associated with recognizing and predicting human movements

Abstract. This article provides a review of the literature and existing research in recent years on the topic, describes the tasks associated with recognizing and predicting human movements, as well as the course of action and detailed steps in preparing the dataset and training models.
Key words: Neural network, physical movement, prediction, robotic image
The development of a set of machine learning methods called deep learning has made it possible to create complex neural network architectures that have sufficient performance and make it possible to solve a wide range of problems in the field of computer vision that could not be effectively solved before. In particular, the problem of recognizing and predicting human actions using machine learning methods is becoming urgently relevant for research in our time. With the development of technology, the number of possible applications is growing: human interaction with technological equipment and robots, productivity analysis and monitoring the safety of workers in an enterprise, recognition systems for video surveillance, the formation of realistic movements in movies and computer games, etc. A fundamental feature of technological progress is the need to invent new ways of processing and analyzing data, as well as methods of interacting with them and adapting them to modern needs. This work is aimed at improving control systems for the use of active or semi-active industrial exoskeletons - devices designed to increase human strength through an external frame. The goal of this article is to create a neural network that can predict the type of human physical activity and the moment of its beginning based on existing data using modern machine learning methods. To achieve the goal, it is necessary to solve the following tasks:
1. Prepare a database based on markup and existing information about movements on video files;
2. Prepare training (dataset) and control samples;
3. Train and test the created neural network.