Human intention understanding method based on deep stacking Bi-LSTM for man-machine cooperation

Abstract

The invention provides a human intention understanding method based on deep stacking Bi-LSTM for man-machine cooperation, and belongs to the technical field of robot control. The method comprises: 1,collecting a motion sequence based on human skeleton node information; step 2, carrying out data preprocessing on the acquired human body motion sequence; step 2, carrying out model training on the deeply stacked bidirectional LSTM network by utilizing the preprocessed data; and step 3, inputting the data processed in the step 2 into a deep stack Bi-LSTM network for training, so as to enable the deep stack Bi-LSTM network to capture data features of a human body motion sequence based on human body skeleton node information; and step 4, testing by using the network model of the deep stack bidirectional LSTM. According to the end-to-end human body intention understanding method based on the motion sequence of the human body skeleton information provided by the invention, the intention understanding effect is better and the precision is more accurate based on the bidirectional LSTM model of depth stacking.

Description

technical field [0001] The invention belongs to the technical field of robot control, and relates to a human-machine cooperation-oriented method for understanding human intentions based on a deep-stacked Bi-directional Long Short-Term Memory (Bi-directional Long Short-Term Memory, Bi-LSTM for short). Background technique [0002] With the improvement of robot intelligence, the development of flexible robot technology, and the increasing demand for cooperation between robots and humans to complete tasks, robot applications are developing in the direction of human-machine collaboration. In order to realize safe and efficient human-machine collaboration, it is necessary for the robot to make timely predictions of human actions, realize the robot's understanding of human behavior intentions and in-depth human-computer interaction, so as to ensure the efficiency of human-machine collaborative operations. Traditional statistics-based methods, such as hidden Markov model (HMM), dyn...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that the robot’s understanding of human behavior intentions is not ideal in the current human-computer collaboration, the present invention proposes a human-computer collaboration-oriented human-machine collaboration-based human intention understanding method based on deep stacked bidirectional LSTM

Images