CN111506190A - Real-time continuous prediction method for human motion intention based on myoelectricity - Google Patents
Real-time continuous prediction method for human motion intention based on myoelectricity Download PDFInfo
- Publication number
- CN111506190A CN111506190A CN202010245346.3A CN202010245346A CN111506190A CN 111506190 A CN111506190 A CN 111506190A CN 202010245346 A CN202010245346 A CN 202010245346A CN 111506190 A CN111506190 A CN 111506190A
- Authority
- CN
- China
- Prior art keywords
- angle
- real
- time
- knee joint
- signals
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Dermatology (AREA)
- Neurology (AREA)
- Neurosurgery (AREA)
- Human Computer Interaction (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to a real-time continuous prediction method of human motion intention based on electromyography, which comprises the steps of extracting EMG signals and IMU signals of single-side lower limbs of a human body, setting an evaluation threshold, selecting the IMU signals acquired by any one of electrodes of thighs and shanks respectively, obtaining a knee joint angle at the current moment through a lower limb kinematics calculation algorithm, resampling the EMG signals of the lower limbs and the knee joint angle at the current moment by adopting a sliding time window method, establishing L STM angle predictors with the same network structure, comparing the angle predictors with the real-time knee joint angle to obtain an evaluation standard root mean square error RMSE, and when the evaluation standard root mean square error RMSE is smaller than the set threshold, outputting the angle of the predictors is accurate.
Description
Technical Field
The invention relates to the technical field of real-time continuous prediction of human movement intentions, in particular to a real-time continuous prediction method of human movement intentions based on myoelectricity.
Background
The robot is a highly fused product of disciplines in various fields, and productivity is greatly liberated since birth. Auxiliary robots, namely exoskeleton robots, power artificial limbs and the like, have great research and development values in the fields of old-age and disabled-assisting, military and the like. The exoskeleton robot is unique in that the working space of the exoskeleton robot is overlapped with the height of a human body, and the robot body needs to be highly cooperated with the human body. When the running speed or the running track of the exoskeleton mechanical arm deviates from the human motion intention, the normal motion of the human body is easily obstructed and even the human body is easily injured, so that the human-computer interaction based on the motion intention perception is required to have higher precision and a reasonable assistance strategy. The traditional exoskeleton becomes a passive exoskeleton which can only passively receive external control instructions and cannot actively judge the assistance time. Human-computer Interaction (HRI) technology enables the exoskeleton to transition from passive mode to active mode to understand the movement intention of the Human body through processing the relevant movement information of the Human body, and performs appropriate assistance according to the movement intention of the Human body.
At present, the bioelectric signals which are concerned about mainly include electroencephalogram (EEG), Electrooculography (EOG) and Electromyography (EMG), compared with other two types of electrical signals, the EMG signals have the characteristics of high correlation, relative stability and the like, and have the convenient conditions of surface noninvasive acquisition and the like, and the bioelectric signals are widely applied to the related fields of medical diagnosis, motion analysis, interactive games and the like.
The EMG signals are used for sensing the human behavior intention, and the human behavior intention sensing device has an important seat in the field of human-computer interaction research at home and abroad. The characteristic that the human body has electromyography-mechanical delay phenomenon, namely the EMG signal is earlier than the action, the advance time is about 25-125ms, and the characteristic provides a possibility for predicting the human body joint kinematic parameters before the action occurs. Brantley et al in 2017 suggested that continuous EMG signals could be used as an estimate of lower limb knee and ankle kinematics parameters. In 2014, Jimson et al introduce the EMD into a reference item when estimating the angle of the finger joint based on EMG time domain characteristics and signal envelopes, and after the result display is introduced into the EMD, the estimation precision of the joint angle is improved. In the country, in 2015, based on a muscle model of Hill, korea jianda et al, who is the shenyang automation research institute, angle estimation is performed on upper limb movement by means of an IMU sensor by only utilizing the wavelength and zero crossing rate characteristics of a single-channel EMG signal of biceps brachii muscle, the fitting effect is very good in closed-loop estimation, and the angle tends to diverge in open-loop estimation. In 2017, Xianceng et al, Shanghai university of transportation, used the EMG signals of the biceps brachii, the triceps brachii, the anterior deltoid, the posterior deltoid and the middle deltoid, and used a combined model of a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN), to estimate the motion trajectory of the upper limbs of the human body, thereby achieving a better effect.
In the process of processing the EMG signal, both the extraction of the robust features and the running of the related programs take time, and the EMG signal analysis result corresponding to the sampling time cannot be well matched with the motion state at the result output time. In addition, mechanical and control delays in the mobility aid itself can exacerbate the problem.
Disclosure of Invention
The invention provides a real-time continuous prediction method of human motion intention based on electromyography, aiming at solving the problems that an EMG signal analysis result cannot be well matched with a motion state at a result output moment, and mechanical hysteresis and control hysteresis exist, and the invention provides the following technical scheme:
a real-time continuous prediction method of human motion intention based on myoelectricity comprises the following steps;
step 1: extracting EMG signals and IMU signals of the lower limbs on one side of the human body, and setting an evaluation threshold value;
step 2: selecting IMU signals acquired by any one of the thigh and the shank respectively, and solving the knee joint angle at the current moment through a lower limb kinematic solution algorithm;
and step 3: resampling the lower limb EMG signals and the knee joint angle at the current moment by adopting a sliding time window method;
step 4, establishing L STM angle predictors with the same network structure and corresponding to five motion modes, putting the EMG signals and the knee joint angles at the current moment into the corresponding L STM angle predictors for training according to the marked motion mode labels, and taking knee joint angle vectors with the same data length after the time stamps are pushed backwards for a certain time as predictor training labels;
step 5, determining a current motion mode through a motion mode classification algorithm according to a trained L STM angle predictor, and inputting data to a corresponding L STM predictor according to the current motion mode to perform real-time angle prediction;
step 6: and pushing the whole time stamp of the output angle value of the predictor forward for a certain time, comparing the time stamp with the real-time knee joint angle to obtain an evaluation standard root mean square error RMSE, and when the evaluation standard root mean square error RMSE is smaller than a set threshold value, accurately outputting the angle by the predictor.
Preferably, the step 1 specifically comprises:
attaching signal acquisition equipment to eight muscles of rectus femoris, vastus lateralis, vastus medialis, tibialis anterior, semitendinosus, biceps femoris longhead, lateral gastrocnemius and medial gastrocnemius on one side of a human body to extract EMG signals and IMU signals, and setting an evaluation threshold;
and recording videos, and judging a motion mode label corresponding to each data point through the videos, wherein the motion modes are divided into five modes of flat ground, ascending stairs, descending stairs, ascending slopes and descending slopes.
Preferably, the signal acquisition equipment adopts a Delsys Trigno electromyography acquisition system, the system comprises a wireless communication base station and 16 wireless electromyography electrodes, and an EMG and IMU signal acquisition device is arranged in each electrode.
Preferably, the step 2 specifically comprises:
the method comprises the steps of selecting IMU signals collected by any one of electrodes of a thigh and a shank respectively, transmitting the collected IMU signals to a development board card through a wireless module when the collected IMU signals are collected for a fixed length, and obtaining the knee joint angle at the current moment by a lower limb kinematics calculation algorithm carried on the development board card, wherein the knee joint angle frequency at the current moment is consistent with the IMU signal frequency.
Preferably, both ends of the sliding time window correspond to synchronous points of the EMG signals and the joint angle signals, the stepping length of the sliding time window ensures that the EMG and angle data in the sliding time window are mutually synchronous, and the lower limb EMG signals and the knee joint angle at the current moment are resampled by adopting a sliding time window method.
Preferably, L STM angle predictors with the same network structure are established and correspond to five motion modes, EMG signals and knee joint angles at the current moment are placed into the corresponding L STM angle predictors to be trained according to marked motion mode labels, knee joint angle vectors with the same data length after time stamps are pushed backwards for 27ms-300ms are used as predictor training labels, a k-fold cross verification method is adopted, one sample is taken as a test set each time, the rest k-1 samples are used as training sets, and k times of experiments are repeated, so that each sample can be used as one test set number.
Preferably, the training data is randomly shuffled in units of samples before each training is started.
Preferably, the step 5 is specifically that the EMG signals and the IMU signals are transmitted to a development board card through a wireless module every time fixed-length data are collected, a motion mode classifier carried on the board card is used for judging and storing a motion mode according to the EMG signals and the IMU signals, a lower limb kinematic calculation algorithm carried on the development board card is used for solving and storing a real-time knee joint angle, the real-time knee joint angle obtained through calculation and the corresponding EMG signals are input to a corresponding L STM angle predictor through the classified motion mode for angle prediction, and a result is stored.
Preferably, the step 6 is specifically that a prediction result of an L STM angle predictor is extracted, a lower limb kinematic solution algorithm is used for solving the real-time knee joint angle, a L STM angle predictor output angle value time stamp is pushed forwards for 27ms-300ms integrally, the real-time knee joint angle is solved with the lower limb kinematic solution algorithm, an evaluation standard root mean square error RMSE is obtained, and the evaluation standard mean square error is expressed by the following formula:
wherein, thetapFor the prediction result of L STM angle predictor, thetarRepresenting the real-time knee joint angle, wherein n is the test frequency;
when the evaluation standard root mean square error RMSE is smaller than a set threshold, the output angle of the predictor is accurate; and when the evaluation standard root mean square error RMSE is larger than a set threshold value, acquiring data again for training.
Preferably, the development board loads an NVIDIA Pascal architecture GPU, 2 Denver 64-bit CPUs, and a quad-core a57 composite processor, and is equipped with a L inux Ubuntu operating system and a communication interface.
The invention has the following beneficial effects:
the training data of the invention are all randomly shuffled (shuffle) by taking samples as units, which can reduce the overfitting influence of fixed data sequence on the parameters of the prediction model, L STM screens related information by setting a 'gate structure' to solve the long-term dependence problem in the recurrent neural network, and simultaneously eliminates the problems of gradient explosion and gradient disappearance in the training process of the recurrent neural network model to a great extent.
Drawings
FIG. 1 is a flow chart of a real-time continuous human movement intention prediction method based on myoelectricity;
FIG. 2 is a schematic view of a signal acquisition device in a mounted position;
FIG. 3 is a validation experiment environment;
FIG. 4 is a lower limb joint angle prediction model network constructed based on L STM.
Detailed Description
The present invention will be described in detail with reference to specific examples.
The first embodiment is as follows:
as shown in FIG. 1, the invention provides a real-time continuous prediction method of human body movement intention based on myoelectricity, which comprises the following steps:
a real-time continuous prediction method of human motion intention based on myoelectricity comprises the following steps;
step 1: extracting EMG signals and IMU signals of the lower limbs on one side of the human body, and setting an evaluation threshold value; the step 1 specifically comprises the following steps:
according to the figure 2, the signal acquisition equipment is attached to eight muscles of rectus femoris, vastus lateralis, vastus medialis, tibialis anterior, semitendinosus, biceps femoris longhead, lateral gastrocnemius and medial gastrocnemius on one side of a human body to extract EMG signals and IMU signals, and an evaluation threshold value is set;
and recording videos, and judging the motion mode labels corresponding to each data point through the videos, wherein the motion modes are divided into five modes of flat ground, ascending stairs, descending stairs, ascending slopes or descending slopes.
According to the invention, as shown in fig. 3, for five lower limb movement modes, namely, flat ground, ascending stairs, descending stairs, ascending slopes and descending slopes, knee joint angle prediction models for different movement modes are developed based on a long-term and short-term memory neural network, so that future knee joint angles can be predicted based on EMG signals and knee joint angles at the current moment, and the lower limb movement intention of a human body can be sensed and predicted at a quantitative analysis angle.
A L ong Short-term Memory neural network (L STM) is one of the variants of the recurrent neural network, has mature application in the fields of speech recognition and language translation and is very suitable for processing time sequence information, L STM screens related information by setting a gate structure to solve the problem of long-term dependence in the recurrent neural network and greatly eliminate the problems of gradient explosion and gradient disappearance in the training process of the recurrent neural network model
The signal acquisition equipment adopts a Delsys Trigno electromyography acquisition system, the system comprises a wireless communication base station and 16 wireless electromyography electrodes, and an EMG and IMU signal acquisition device is arranged in each electrode.
Step 2: selecting IMU signals acquired by any one of the thigh and the shank respectively, and solving the knee joint angle at the current moment through a lower limb kinematic solution algorithm;
the step 2 specifically comprises the following steps:
the method comprises the steps of selecting IMU signals acquired by any one of thigh and shank electrodes respectively, transmitting acquired fixed-length data of the IMU signals to an NVIDIA Jetson TX2 development board card through a wireless module, obtaining a knee joint angle at the current moment by a lower limb kinematics calculation algorithm carried on the NVIDIA Jetson TX2 development board card, wherein the knee joint angle frequency at the current moment is consistent with the IMU signal frequency, carrying an NVIDIA Pascal architecture GPU, 2 Denver 64-bit CPUs and a quad-core A57 composite processor on the NVIDIA Jetson TX2 development board card, and matching with a L inux Ubuntu operating system and a communication interface.
And step 3: resampling the lower limb EMG signals and the knee joint angle at the current moment by adopting a sliding time window method; the two ends of the sliding time window correspond to synchronous points of the EMG signals and the joint angle signals, the stepping length of the sliding time window ensures that the EMG and angle data in the sliding time window are mutually synchronous, and the lower limb EMG signals and the knee joint angle at the current moment are resampled by adopting a sliding time window method.
The method comprises the steps of step 4, establishing L STM angle predictors with the same network structure and corresponding to five motion modes, putting EMG signals and knee joint angles at the current time into corresponding L STM angle predictors to be trained according to marked motion mode labels, taking knee joint angle vectors with the same data length after a time stamp is pushed backwards for 27-300ms as predictor training labels, establishing L STM angle predictors with the same network structure and corresponding to the five motion modes, putting the EMG signals and the knee joint angles at the current time into corresponding L STM angle predictors to be trained according to the marked motion mode labels, taking the knee joint angle vectors with the same data length after the time stamp is pushed backwards for 27-300ms as predictor training labels, adopting a k-fold cross-validation method, taking one sample as a test set each time, taking the rest k-1 samples as training sets, repeating k times of experiments, enabling each sample to serve as a test set number, and carrying out random data by taking the sample as a unit before each training begins, so that the influence of fixed data sequence on shuffle on prediction model parameters can be reduced.
Step 5, determining a current motion mode of the trained L STM angle predictor through a motion mode classification algorithm, and inputting data to the corresponding L STM predictor according to the current motion mode to perform real-time angle prediction;
the step 5 specifically includes the steps that each time the EMG signal and the IMU signal are collected, fixed-length data are transmitted to an NVIDIA Jetson TX2 development board card through a wireless module, a motion mode classifier carried on the NVIDIA Jetson TX2 board card is used for judging and storing a motion mode according to the EMG and IMU signals, a lower limb kinematics calculation algorithm carried on the NVIDIA Jetson TX2 development board card is used for calculating and storing a real-time knee joint angle, the calculated real-time knee joint angle and the corresponding EMG signal are input to a corresponding STM L angle predictor through the classified motion mode for angle prediction, and a result is stored.
Step 6: and pushing the whole time stamp of the output angle value of the predictor forwards for 27-300ms, comparing the time stamp with the real-time knee joint angle to obtain an evaluation standard root mean square error RMSE, and when the evaluation standard root mean square error RMSE is smaller than a set threshold value, the output angle of the predictor is accurate.
The step 6 is specifically that a prediction result of an L STM angle predictor is extracted, a lower limb kinematic solution algorithm is used for solving the real-time knee joint angle, a L STM angle predictor output angle value timestamp is pushed forwards by 27ms-300ms integrally, the real-time knee joint angle is solved with the lower limb kinematic solution algorithm, an evaluation standard root mean square error RMSE is obtained, and the evaluation standard mean square error is expressed by the following formula:
wherein, thetapFor the prediction result of L STM angle predictor, thetarRepresenting the real-time knee joint angle, n ═ k;
when the evaluation standard root mean square error RMSE is smaller than a set threshold, the output angle of the predictor is accurate; and when the evaluation standard root mean square error RMSE is larger than a set threshold value, acquiring data again for training.
The NVIDIA Jetson TX2 development board card carries an NVIDIA Pascal architecture GPU, 2 Denver 64-bit CPUs and a quad-core A57 composite processor, and is provided with a L inux Ubuntu operating system and a communication interface.
The above description is only a preferred embodiment of the real-time continuous human movement intention prediction method based on myoelectricity, and the protection scope of the real-time continuous human movement intention prediction method based on myoelectricity is not limited to the above embodiments, and all technical solutions belonging to the idea belong to the protection scope of the present invention. It should be noted that modifications and variations which do not depart from the gist of the invention will be those skilled in the art to which the invention pertains and which are intended to be within the scope of the invention.
Claims (10)
1. A real-time continuous prediction method of human motion intention based on myoelectricity is characterized in that: comprises the following steps;
step 1: extracting EMG signals and IMU signals of the lower limbs on one side of the human body, and setting an evaluation threshold value;
step 2: selecting IMU signals acquired by any one of the thigh and the shank respectively, and solving the knee joint angle at the current moment through a lower limb kinematic solution algorithm;
and step 3: resampling the lower limb EMG signals and the knee joint angle at the current moment by adopting a sliding time window method;
step 4, establishing L STM angle predictors with the same network structure and corresponding to five motion modes, putting the EMG signals and the knee joint angles at the current moment into the corresponding L STM angle predictors for training according to the marked motion mode labels, and taking knee joint angle vectors with the same data length after the time stamps are pushed backwards for a certain time as predictor training labels;
step 5, determining a current motion mode through a motion mode classification algorithm according to a trained L STM angle predictor, and inputting data to a corresponding L STM predictor according to the current motion mode to perform real-time angle prediction;
step 6: and pushing the whole time stamp of the output angle value of the predictor forward for a certain time, comparing the time stamp with the real-time knee joint angle to obtain an evaluation standard root mean square error RMSE, and when the evaluation standard root mean square error RMSE is smaller than a set threshold value, accurately outputting the angle by the predictor.
2. The myoelectricity-based human movement intention real-time continuous prediction method according to claim 1, which is characterized in that: the step 1 specifically comprises the following steps:
attaching signal acquisition equipment to eight muscles of rectus femoris, vastus lateralis, vastus medialis, tibialis anterior, semitendinosus, biceps femoris longhead, lateral gastrocnemius and medial gastrocnemius on one side of a human body to extract EMG signals and IMU signals, and setting an evaluation threshold;
and recording videos, and judging a motion mode label corresponding to each data point through the videos, wherein the motion modes are divided into five modes of flat ground, ascending stairs, descending stairs, ascending slopes and descending slopes.
3. The myoelectricity-based human movement intention real-time continuous prediction method according to claim 2, which is characterized in that: the signal acquisition equipment adopts a Delsys Trigno electromyography acquisition system, the system comprises a wireless communication base station and 16 wireless electromyography electrodes, and an EMG and IMU signal acquisition device is arranged in each electrode.
4. The myoelectricity-based human movement intention real-time continuous prediction method according to claim 1, which is characterized in that: the step 2 specifically comprises the following steps:
the method comprises the steps of selecting IMU signals collected by any one of electrodes of a thigh and a shank respectively, transmitting the collected IMU signals to a development board card through a wireless module when the collected IMU signals are collected for a fixed length, and obtaining the knee joint angle at the current moment by a lower limb kinematics calculation algorithm carried on the development board card, wherein the knee joint angle frequency at the current moment is consistent with the IMU signal frequency.
5. The myoelectricity-based human movement intention real-time continuous prediction method according to claim 1, which is characterized in that: the two ends of the sliding time window correspond to synchronous points of the EMG signals and the joint angle signals, the stepping length of the sliding time window ensures that the EMG and angle data in the sliding time window are mutually synchronous, and the lower limb EMG signals and the knee joint angle at the current moment are resampled by adopting a sliding time window method.
6. The real-time continuous human body movement intention prediction method based on the electromyography as claimed in claim 1 is characterized in that L STM angle predictors with the same network structure are established and correspond to five movement modes, EMG signals and knee joint angles at the current moment are placed into the corresponding L STM angle predictors to be trained according to marked movement mode labels, knee joint angle vectors with the same data length after a time stamp is pushed backwards for 27ms-300ms are used as predictor training labels, a k-fold cross verification method is adopted, one sample is taken as a test set each time, the rest k-1 samples are used as training sets, and k times of experiments are repeated, so that each sample can be used as one test set number.
7. The real-time continuous myoelectricity-based human body movement intention prediction method according to claim 6, which is characterized in that: the training data is randomly shuffled in sample units before each training session begins.
8. The real-time continuous human body movement intention prediction method based on the electromyography of claim 1 is characterized in that the step 5 specifically comprises the steps of transmitting the EMG signal and the IMU signal to a development board card through a wireless module every time the fixed-length data is acquired, judging and storing a movement mode according to the EMG signal and the IMU signal through a movement mode classifier carried on the board card, solving and storing a real-time knee joint angle through a lower limb kinematics calculation algorithm carried on the development board card, inputting the calculated real-time knee joint angle and the corresponding EMG signal into a corresponding L STM angle predictor through the classified movement mode to carry out angle prediction, and storing a result.
9. The real-time continuous human movement intention prediction method based on myoelectricity as claimed in claim 1, wherein the step 6 is specifically that the prediction result of L STM angle predictor is extracted and the lower limb kinematics calculation algorithm is used for solving the real-time knee joint angle, the L STM angle predictor output angle value time stamp is pushed forwards by 27ms-300ms integrally, and the real-time knee joint angle is solved with the lower limb kinematics calculation algorithm to obtain the evaluation standard root mean square error RMSE, and the evaluation standard mean square error is expressed by the following formula:
wherein, thetapFor the prediction result of L STM angle predictor, thetarRepresenting the real-time knee joint angle, wherein n is the test frequency;
when the evaluation standard root mean square error RMSE is smaller than a set threshold, the output angle of the predictor is accurate; and when the evaluation standard root mean square error RMSE is larger than a set threshold value, acquiring data again for training.
10. The myoelectricity-based human movement intention real-time continuous prediction method according to claim 4 or 8, wherein the development board card is loaded with an NVIDIA Pascal architecture GPU, 2 Denver 64-bit CPUs and a quad-core A57 composite processor, and is provided with a L inux Ubuntu operating system and a communication interface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010245346.3A CN111506190A (en) | 2020-03-31 | 2020-03-31 | Real-time continuous prediction method for human motion intention based on myoelectricity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010245346.3A CN111506190A (en) | 2020-03-31 | 2020-03-31 | Real-time continuous prediction method for human motion intention based on myoelectricity |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111506190A true CN111506190A (en) | 2020-08-07 |
Family
ID=71864769
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010245346.3A Pending CN111506190A (en) | 2020-03-31 | 2020-03-31 | Real-time continuous prediction method for human motion intention based on myoelectricity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111506190A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112434630A (en) * | 2020-12-01 | 2021-03-02 | 深圳先进技术研究院 | Training method of continuous motion information prediction model and training device and equipment thereof |
| CN113534960A (en) * | 2021-07-29 | 2021-10-22 | 中国科学技术大学 | Upper arm prosthesis control method and system based on IMU and surface electromyographic signals |
| CN114767463A (en) * | 2022-03-11 | 2022-07-22 | 上海电机学院 | Consciousness control exercise rehabilitation system and method based on surface myoelectricity |
| CN115399791A (en) * | 2022-06-28 | 2022-11-29 | 天津大学 | Stroke lower limb function assessment method and system based on myoelectric motion multi-data fusion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180307314A1 (en) * | 2012-06-14 | 2018-10-25 | Medibotics Llc | Circumferential Array of Electromyographic (EMG) Sensors |
| CN109657651A (en) * | 2019-01-16 | 2019-04-19 | 杭州电子科技大学 | A kind of continuous method for estimating of lower limb knee joint based on electromyography signal |
| CN109940584A (en) * | 2019-03-25 | 2019-06-28 | 杭州程天科技发展有限公司 | The detection method that a kind of exoskeleton robot and its detection human motion are intended to |
-
2020
- 2020-03-31 CN CN202010245346.3A patent/CN111506190A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180307314A1 (en) * | 2012-06-14 | 2018-10-25 | Medibotics Llc | Circumferential Array of Electromyographic (EMG) Sensors |
| CN109657651A (en) * | 2019-01-16 | 2019-04-19 | 杭州电子科技大学 | A kind of continuous method for estimating of lower limb knee joint based on electromyography signal |
| CN109940584A (en) * | 2019-03-25 | 2019-06-28 | 杭州程天科技发展有限公司 | The detection method that a kind of exoskeleton robot and its detection human motion are intended to |
Non-Patent Citations (1)
| Title |
|---|
| 郭浩: "基于EMG的人体下肢运动意图感知与预测研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112434630A (en) * | 2020-12-01 | 2021-03-02 | 深圳先进技术研究院 | Training method of continuous motion information prediction model and training device and equipment thereof |
| CN112434630B (en) * | 2020-12-01 | 2022-08-16 | 深圳先进技术研究院 | Training method of continuous motion information prediction model and training device and equipment thereof |
| CN113534960A (en) * | 2021-07-29 | 2021-10-22 | 中国科学技术大学 | Upper arm prosthesis control method and system based on IMU and surface electromyographic signals |
| CN113534960B (en) * | 2021-07-29 | 2024-05-28 | 中国科学技术大学 | Upper arm artificial limb control method and system based on IMU and surface electromyographic signals |
| CN114767463A (en) * | 2022-03-11 | 2022-07-22 | 上海电机学院 | Consciousness control exercise rehabilitation system and method based on surface myoelectricity |
| CN115399791A (en) * | 2022-06-28 | 2022-11-29 | 天津大学 | Stroke lower limb function assessment method and system based on myoelectric motion multi-data fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111506190A (en) | Real-time continuous prediction method for human motion intention based on myoelectricity | |
| CN108785997B (en) | Compliance control method of lower limb rehabilitation robot based on variable admittance | |
| CN111506189B (en) | Motion mode prediction and switching control method for complex motion of human body | |
| KR20150131554A (en) | Human joint kinematics information extraction method from multi-channel surface electromyogram signals, recording medium and device for performing the method | |
| CN110801226A (en) | Human knee joint moment testing system method based on surface electromyographic signals and application | |
| Foroutannia et al. | A deep learning strategy for EMG-based joint position prediction in hip exoskeleton assistive robots | |
| Lu et al. | Continuous and simultaneous estimation of lower limb multi-joint angles from sEMG signals based on stacked convolutional and LSTM models | |
| Lu et al. | Evaluation of classification performance in human lower limb jump phases of signal correlation information and LSTM models | |
| Lu et al. | Effective recognition of human lower limb jump locomotion phases based on multi-sensor information fusion and machine learning | |
| Koch et al. | A recurrent neural network for hand gesture recognition based on accelerometer data | |
| Yi et al. | Smart healthcare-oriented online prediction of lower-limb kinematics and kinetics based on data-driven neural signal decoding | |
| Liang et al. | Synergy-based knee angle estimation using kinematics of thigh | |
| Chen et al. | Gait recognition for lower limb exoskeletons based on interactive information fusion | |
| Tao et al. | Multi-channel sEMG based human lower limb motion intention recognition method | |
| Zhang et al. | ViT-LLMR: Vision Transformer-based lower limb motion recognition from fusion signals of MMG and IMU | |
| Geng et al. | Lower limb joint nursing and rehabilitation system based on intelligent medical treatment | |
| Song et al. | Continuous online prediction of lower limb joints angles based on sEMG signals by deep learning approach | |
| CN109645995A (en) | Joint motions estimation method based on myoelectricity model and Unscented kalman filtering | |
| Baptista et al. | new approach of cycling phases detection to improve FES-pedaling in SCI individuals | |
| Li et al. | Prediction of knee joint moment by surface electromyography of the antagonistic and agonistic muscle pairs | |
| Mironov et al. | Myoelectric control system of lower limb exoskeleton for re-training motion deficiencies | |
| CN115762708A (en) | Lower limb rehabilitation scheme design method and device using multi-source information and case reasoning | |
| Xie et al. | [Retracted] Sensor‐Based Exercise Rehabilitation Robot Training Method | |
| CN115985464B (en) | Muscle fatigue classification method and system based on multi-mode data fusion | |
| Chowdary et al. | Artificial Intelligence-based approach for gait pattern identification using surface electromyography (SEMG) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |