Boosted exoskeleton motion intent and pace planning method based on multi-sensing information

Abstract

The invention provides a boosted exoskeleton motion intent and pace planning method based on multi-sensing information. The method comprises the following steps: S1. acquiring vertical-direction acceleration signals of legs of a human body and oxygen consumption data of the human body, and acquiring surface electromyographic signals from muscles of the legs; S2. taking characteristic parameters ofthe acceleration signals, characteristic parameters of the oxygen consumption data and characteristic parameters of the surface electromyographic signals as inputs of a training model; and S3. establishing an LSTM deep learning network so as to obtain a corresponding relationship between the characteristic parameters of the surface electromyographic signals and a pace plan, a corresponding relationship between the characteristic parameters of the surface electromyographic signals and an exoskeleton boosting grade, and a corresponding relationship between the characteristic parameters of the surface electromyographic signals and a motion mode, and taking the boosting grade, the motion mode and the pace plan as outputs of the model. According to the method provided by the invention, key parameters required by a boosted exoskeleton control system can be comprehensively presented, the resolving efficiency of real-time intents is increased, and rapid and accurate sharing, decision making and executing can be achieved.

Description

technical field [0001] The invention belongs to the technical field of man-machine cooperative control of exoskeleton robots, and in particular relates to a multi-sensing information-based power-assisted exoskeleton motion intention and gait planning method, which is used for power-assisted exoskeleton robot active intention recognition, power-assisted efficiency level classification and Real-time gait planning. Background technique [0002] With the continuous development of science and technology, exoskeleton robots have achieved unprecedented development in the military and civilian fields. The most notable feature of the exoskeleton is the real-time interaction with the wearer throughout the whole process. The wearer has real physical contact with the exoskeleton body, which is a human-machine fusion system. The motion control of exoskeleton robots not only needs to have a high degree of stability and robustness, but also needs to be "natural" and "adaptable". "Natural...

AI Technical Summary

Problems solved by technology

[0004] However, the existing motion intention recognition methods have the following problems: ① The human-computer interaction information from a single motion sensor is too one-sided, and cannot effectively express the motion intention and status of the human-machine system during the interaction process.

②In terms of understanding and prediction of human motion intentions, the existing methods use a single motion sensor to collect data and are multi-oriented to judge the current motion state, and cannot accurately obtain future motion information, so they cannot compensate for errors caused by mechanical delays

③ In terms of application based on surface EMG signals, the existing algorithm structure is too simple to fully extract the future movement intention information contained in EMG

④, using the same behavior in the interaction process of the human-computer system will generate a variety of information, such as surface electromyographic signals, joint angle information, etc., which are very important for accurately estimating the state of human-machine motion and the behavioral intention of the rehabilitation person. However, the real need for exoskeletons What information should be included in the recognition of human motion intentions, but there is no specific concept. Currently, it only targets a specific amount of motion, such as joint angle information, etc.

Therefore, when interpreting the connotation of human intentions, it is necessary to consider not only the calculation efficiency of real-time intentions, but also after obtaining the motion state of the human-machine system and human motion intentions. However, how to achieve fast and accurate sharing, decision-making, and execution is relatively rare.

Images