36 results about "Motor behaviour" patented technology

Method for estimating movement behaviour of a user of a mobile communication device by a neural network comprising one or more lower and one or more higher hidden layers. The method comprising a stepof obtaining (401) sensor data from sensors in the mobile device; a step of obtaining (402) measurements related to a movement of the user; a step of labelling (403) these measurements as weakly labelled data; pre-training (404) the lower hidden layers to estimate the measurements from the first set of sensor data; a step of obtaining (405) a second set of sensor data wherein movement behaviour ofthe user is labelled as labelled data; a step of training (406) the higher hidden layers with the labelled data to estimate the movement behaviour of the user as said output.

A mobile medical device drivable on rollers or wheels on the floor by way of a motor is provided, wherein the medical device is configured to take a planned route. The device has an optical display apparatus embodied to optically display at least one future movement feature of the medical device on the planned route. A visualization of the route planning helps the user to better estimate the future movement behavior of a medical device.

The invention discloses a human body identification method based on a sequence depth image, comprising: collecting the sequence depth image of a human body, and using a convolutional neural network and a recursive neural network to perform regression of human body joint points based on the sequence depth image to obtain the coordinate positions of the human body joint points ; Use the coordinates of the joint points of the human body to calculate the body posture feature vector, use the static feature classifier to identify, and obtain the human body identity recognition score of the static feature; extract the human body motion behavior feature from the sequence depth image, and use the dynamic feature classifier to perform Recognition to obtain the human body recognition score of dynamic features; the weighted calculation of the static and dynamic features of the human body recognition score to obtain the human body recognition score, and then obtain the human body recognition result. The present invention has strong robustness, and the accuracy rate is not affected by application scenarios.

The invention relates to the technical field of motion behavior recognition, in particular to a method and device for animal motion behavior discrimination based on temporal correlation analysis. Obtain the known continuous motion state of the animal and the corresponding three-axis acceleration information, as well as the continuous three-axis acceleration information to be discriminated; convert the three-axis acceleration information into three-dimensional data in the natural coordinate system, and the three-axis acceleration of the known motion state The information and the corresponding three-dimensional data constitute the training data, and the continuous three-axis acceleration information to be judged and the corresponding three-dimensional data constitute the test data; according to the time series correlation relationship between the motion behavior states and the iterative regression tree model obtained through the training data training, the test The data is used to discriminate the state of motion behavior, combined with the timing of the behavior state, it makes up for the shortcomings of the iterative decision tree's poor performance in processing highly relevant data, and further improves the accuracy of behavior discrimination.

The invention belongs to the field of mobile robot formation motion planning, and specifically relates to an improved adaptive zero-space behavior fusion method for a multi-robot formation, comprising the following steps: decomposing the formation motion process of robots into three types of motion behaviors: running towards a target point, avoiding Obstacles, formation; determine the execution order of 3 kinds of motion behaviors, and its priority order is to avoid obstacles, rush to the target point, formation; establish the motion model of each behavior; solve the behavior motion model according to the robot motion information; the method of the present invention Based on the traditional zero-space behavior fusion method, a logic-priority obstacle selection strategy between robots is proposed in the obstacle avoidance behavior, which effectively avoids the "overshooting" caused by double selection when two robots are obstacles to each other. The phenomenon of "avoiding the distance" and deviating from the mission route improves the efficiency of multi-robot formation missions, is conducive to the maintenance of formation formations, and has made significant progress in motion planning.

The invention provides a somatosensory motion simulation system. The simulation system comprises a user gravity center movement data acquisition unit, a standard movement model generation unit, a processing unit, a safety protection unit and a teaching unit. According to the simulation system provided by the invention, the gravity center movement data of a user can be collected in real time, whether the movement of the user is standard or not is analyzed according to the gravity center movement data of the user, when great deviation exists between the movement of the user and the standard movement, prompt is given out or the on-going movement is controlled, so that the movement behavior of the user is corrected and improved, further, the user masters the essential of the movement rapidly,therefore, the safety of the movement is guaranteed, in addition, the system has the functions of timely protection, movement body sensing and motion load generation and management, thus the user canbe helped to overcome the psychological disorder, the waist, back and leg strength is increased, and the training comprehensive effect is promoted.

The invention discloses a lifting table and a control method and a control device thereof. The lifting table comprises a liftable supporting column, a table top mounted on the supporting column; and a plurality of sensors and a control device connected to the sensors, wherein the sensors are used for analyzing the moving direction of the to-be-detected object located in the lifting table detection area so as to obtain the moving track of the to-be-detected object; the control device is used for setting the motion behavior of at least one to-be-tested object and the control action of the lifting table corresponding to each motion behavior, and is also used for acquiring the motion trail of the to-be-tested object, judging whether the motion trail conforms to the motion behavior or not, and controlling the lifting table to perform corresponding control action if the motion trail conforms to the motion behavior. According to the lifting table and the control method and device thereof provided by the embodiment of the invention, a user can control the lifting table by adopting gesture actions more conforming to own operation habits, the requirements of the user are met, and the user experience is improved.

The present invention discloses a method and system for recognizing the movement behavior of a legged robot based on multi-modal perception, wherein the method includes: based on various types of sensors, collecting information obtained by the legged robot repeatedly walking a preset distance on different types of ground The generated multi-modal sensing data information; the multi-modal sensing data information is divided into data sets according to different types of ground to obtain multi-modal sensing data sets of different ground types; multi-modal sensing data sets based on different ground types The sense data set trains the hidden Markov model of the viscous hierarchical Dirichlet process to obtain the hidden Markov model of the optimal viscous hierarchical Dirichlet process; input the information of the samples to be identified into the hidden Markov model of the optimal viscous hierarchical Dirichlet process The husband model is calculated, and the sum of the logarithmic likelihood function values ​​is calculated; based on the calculation results, the movement behavior recognition of the legged robot is carried out. In the implementation of the present invention, the reliability and accuracy of motion behavior recognition are improved.

The invention discloses a construction method of a video detection model of Parkinson's disease motion retardation based on a deep neural network, which combines computer vision and deep learning technology, evaluates the motion pattern of a test subject through short videos, and judges that there is motion Possibility of delayed symptoms. The present invention considers both the movement behavior and the movement process, proposes a detailed data collection method and a movement trajectory definition method, and designs three novel metrics and a periodic movement network model PMNet to deal with the problem of bradykinesia symptom judgment. Different from the traditional method, the present invention has the characteristics of scalability and portability. The key point extraction method based on the convolutional neural network can be replaced by other more accurate models, and more features can be added to describe the motion behavior. In addition, the present invention is capable of performing other similar periodic motion assessments, such as finger tapping actions of item 3.4 of the MDS-UPDRS scale.

This application provides a robot motion behavior control scheme, which sets a motion behavior task stack constructed according to the principle of first-in-last-out, which is used to store each motion behavior planning module that controls the robot's behavior, and iteratively traverses all processing tasks in an iterative manner. The motion behavior planning module in the motion behavior task stack, so that each motion behavior planning module in it is driven by input information to perform collaborative processing until the motion behavior task stack is empty, thereby realizing the motion behavior control of the robot. As a result, the control process of motion behavior can be reused and expanded in a modular and hierarchical manner, and the same motion behavior planning module can be called with different logic in different control tasks, which is conducive to improving the versatility of the scheme. Reduce development difficulty.

The invention relates to the field of multi-robot formation control, in particular to a multi-robot formation adaptive zero-space behavior fusion method. According to the robot formation movement intention, the robot movement process is decomposed, the behavior sequence steps are determined, and then the movement model steps are established, and the execution sequence of the three movement behaviors is solved to obtain three movement models. Motion information, solve to get the gain coefficient, and fuse the motion behavior steps to get the final speed and direction. Compared with the traditional zero-space behavior fusion method, the present invention makes the solved speed not changed by the working condition environment, has good adaptability, and takes into account both efficiency and performance, and can effectively control the speed. with significant progress.