39 results about "Interaction field" patented technology

The invention discloses a human-computer interaction method and system based on a vision-language-action model, and belongs to the field of human-computer interaction. According to the method, an anchoring ring strategy is adopted to collect human teaching data to finely adjust the VLA model, and then the trained model is applied to an actual human-computer interaction scene. In the data acquisition stage, a first operator guides a master robot to execute task actions, and a slave robot synchronously moves and interacts with a second operator, and returns to a predefined initial position after each interaction; a teaching sample is formed by recording a robot state, an environment image and an instruction text, and a high-quality data set is generated through data enhancement. In the application stage, the real-time robot state, the environment image and the instruction text serve as input, an action instruction is generated through the VLA model, and the robot is driven to complete a cooperation task. According to the method, the data utilization efficiency and the model generalization ability are remarkably improved, the difference between simulation and the real environment is effectively overcome, and efficient, safe and natural man-machine cooperation is achieved.

The invention is applicable to the technical field of electroencephalogram emotion recognition, and provides an electroencephalogram emotion recognition method based on deep learning, which comprises the following steps: firstly, preprocessing electroencephalogram signals, including baseline noise removal, standardization, band-pass filtering and signal segmentation; then establishing a deep learning model, wherein the deep learning model sequentially comprises an automatic encoder, a dynamic graph convolutional neural network, a Transform and a classifier; training a deep learning model by using a cross entropy loss function in combination with regularization; and finally, performing emotion recognition on the preprocessed electroencephalogram signals based on the trained model, and outputting a result. According to the method, multiple deep learning models are fused, so that collaborative extraction of the dynamic spatial features and the long-range time dependence features in the electroencephalogram signals is realized. According to the method, the accuracy of electroencephalogram emotion recognition is greatly improved, and a more reliable and efficient emotion recognition technical support is provided for the man-machine interaction fields such as depression evaluation and affective disorder treatment.

The invention discloses a multi-modal emotion calculation method and system, and mainly relates to the technical field of artificial intelligence and human-computer interaction. Comprising the following steps: synchronously collecting voice, visual and tactile multi-modal data of a user, and carrying out feature extraction on each modal data; performing cross-modal fusion on the extracted multi-modal features, including time sequence alignment and semantic association modeling, and generating a global emotion feature vector; performing dynamic emotion reasoning based on the global emotion feature vector, and outputting an emotion category and emotion intensity; generating a tactile feedback signal according to the emotion category and the emotion intensity, and driving an actuator to output; and updating the emotion memory graph based on user interaction data to complete personalized model self-adaption. The method has the beneficial effects that high-precision and low-delay emotion recognition and natural tactile feedback in a cross-culture scene are realized, and the core bottleneck of dynamic drift adaptation failure and emotion-behavior feedback decoupling in the prior art is solved.

The invention discloses a flexible pressure sensor based on ionized hydrogel and a preparation method thereof. The sensor adopts a sandwich layered structure of a flexible composite electrode, an ionized water gel dielectric layer and a flexible composite electrode, ionized water gel is used as a dielectric layer, and an upper flexible electrode and a lower flexible electrode are prepared from a multi-walled carbon nanotube / polydimethylsiloxane (MWCNTs / PDMS) composite material; the surface, making contact with the dielectric layer, of the electrode is copied through an abrasive paper template method to form a micron-sized rough structure. Under the action of external pressure, the effective contact area between the microstructure electrode and the ionized water gel dielectric layer and the equivalent electrode distance change, so that the interface electric double-layer capacitance is regulated and controlled, and high-sensitivity conversion from a pressure signal to a capacitance signal is realized. The flexible pressure sensor is simple in structure and excellent in flexibility, can be further integrated into a pressure sensing array, and is suitable for the fields of wearable electronics, electronic skin, man-machine interaction and the like.

The application discloses a kind of based on directional shuttle-shaped cavity bimodal flexible capacitive sensor and preparation method thereof, belong to flexible electronic sensor technical field, by dispersing magnetic nanoparticles in low boiling point solvent to form magnetic fluid, emulsification dispersion in hydrogel prepolymer liquid, after rheological control, in turn carry out uniform magnetic field pre-stretching and pulse magnetic field strengthening lengthening, make droplet form shuttle-shaped;Again by half freezing lock shape, vacuum freeze drying removes solvent and makes magnetic nanoparticles deposit in cavity inner wall, completely freeze polymerization makes matrix solidification, finally in situ reduction forms conductive layer and is packaged and is prepared sensor, it has bimodal: when pressure is along the long axis direction of shuttle-shaped cavity, it is first electric capacity change rate, when it is along short axis direction, it is second electric capacity change rate, and second electric capacity change rate is greater than first electric capacity change rate, and pressure direction is identified by single electric capacity pair.The application structure is simple, sensitivity is high, stability is good, is suitable for electronic skin, intelligent prosthesis and man-machine interaction field.

The embodiment of the application discloses a kind of display method, device, terminal, storage medium and product of virtual stack, belong to man-machine interaction field.The method includes: display virtual material in virtual environment, virtual material is the material that virtual object falls;Control first virtual object and move virtual material to target position in virtual environment, target position is the virtual material storage position of the camp that first virtual object is in;Display updated first virtual stack, first virtual stack is the virtual stack that the camp that first virtual object is in has, and virtual stack is obtained by virtual material construction.In the embodiment of the application, by providing the mode of constructing virtual stack, to enrich the mode of battle in virtual environment, avoid single battle mode, help to improve man-machine interaction effect.

The invention provides an end-to-end dialogue system data processing method and a related device, and relates to the field of man-machine interaction.The method comprises the steps that a large model and a small model are arranged in an end-to-end interaction system, and when whether the large model can be called in the historical interaction field or not, the large model is called to respond to an obtained corpus; otherwise, the small models are adopted to respond to the obtained corpora, better end-to-end dialogue experience is achieved in the mode that the large models and the small models are fused, the specialty of the small models is reserved, and the expansibility of the large models is well achieved.

Embodiments of the present application provide a kind of action sequence execution method, electronic equipment, storage medium and program product.It relates to human-computer interaction field, the method includes: based on the semantic feature of representing user instruction, generate global action sequence, global action sequence includes multiple node actions with execution order in it;After executing each node action, each node action is identified, and the identification result corresponding to each node action is obtained;In the case where the identification result indicates that the target node action exists exception, generate the target action sequence with the target node action as starting point;Target action sequence is used to update global action sequence;According to the updated global action sequence, the interface element in operation interface is executed to process until the execution target of user instruction is completed.In the technical scheme of the present application, the technical problem that the success rate of automation execution process is not ideal in the related art is solved.

The invention discloses an intelligent multimedia man-machine interaction system, and relates to the man-machine interaction field, the intelligent multimedia man-machine interaction system comprises an interaction rocking bar and a feedback platform, the interaction rocking bar is in signal connection with the feedback platform, the top of the interaction rocking bar is provided with a control button, and the feedback platform uses an information display system. Comprising an information compression module, an information display module and a custom module, and the information compression module uses a lossy compression algorithm to store multimedia information. According to the method, the keywords are analyzed according to the input content, the corresponding information is retrieved in the multimedia database according to the keywords, the effect of improving the accuracy of retrieved data is achieved, the error range is reduced, meanwhile, the framework layer establishes the relation between the keywords and the multimedia information and builds an information framework, and the information framework is more accurate. The information retrieved according to the keyword covers an information frame established according to the multimedia information to form a main interface and an auxiliary interface, and the main interface and the auxiliary interface are displayed on a feedback platform after being processed by a display layer.

The application discloses a kind of exploration method, device, equipment, medium and program product in virtual world, belong to man-machine interaction field.The method comprises: display is located in the master virtual character in the virtual world, the master virtual character has pet virtual character for exploration;In response to first instruction for the pet virtual character, the pet virtual character is controlled to be attached on the limb of the master virtual character;In response to the master virtual character enters the state of aiming, the enemy virtual character in the preset range of the master virtual character is explored;In response to the enemy virtual character in the preset range of the master virtual character, the position information of the enemy virtual character is displayed.The application provides a new exploration mode, can assist user to explore the position information of enemy virtual character, improves man-machine interaction efficiency.

The invention discloses a user behavior anomaly detection method and system oriented to computing network transactions, and relates to the technical field of user detection oriented to computing network transactions, and the method comprises the following steps: S1, carrying out cross-class recording multi-dimensional fusion: obtaining a user behavior record of a system, the behavior record comprising multi-dimensional data, based on a preset two-dimensional feature window, performing sliding division on the behavior record to obtain a plurality of cross-class sub-records; and dividing the cross-class sub-records on an interaction domain and a numerical domain, projecting the divided sub-records to a multi-domain space formed by the interaction domain and the numerical domain, and generating a cross-class multi-domain space representation to fuse cross-class feature information. According to the method, the cross-class and cross-domain information processing capability is improved, the adaptability of a dynamic scene and heterogeneous distribution is enhanced, and the cooperation efficiency and the detection reliability under the limited bandwidth are improved.

The invention discloses a fabric hybrid electronic integrated system with a stretchable interface. The fabric hybrid electronic integrated system comprises a laser modified fabric substrate, a rigid chip, a printed electrode and a printed interconnection wire. The laser modified fabric is obtained by irradiating a photosensitive polymer composite fabric with laser, can be designed in a customized manner and manufactured in one step by regulating and controlling laser parameters, and simultaneously realizes three different effects: (1) the interface bonding of a printed layer is enhanced; (2) the cross-layer short circuit of the interconnected wires is inhibited; and (3) strain isolation among the rigid chip, the elastic fabric and the interconnected wires. The mechanically enhanced and surface modified fabric realizes heterogeneous integration of a rigid chip and a circuit on the fabric, has interface stretchability, and has potential in the fields of wearable health management and human-computer interaction.

The invention provides an environment transaction alarm system, which relates to the technical field of environment transaction alarm, and is characterized in that multi-source fluctuation data of an underground environment is acquired in real time, and a three-dimensional dynamic model of the underground environment is constructed through the multi-source fluctuation data; determining a three-dimensional dynamic molecular matrix of the underground environment according to the three-dimensional dynamic model of the underground environment, and determining a dynamic interaction field domain of the underground environment through the three-dimensional dynamic molecular matrix; establishing an underground environment transaction monitoring model according to the dynamic interaction field domain of the underground environment, performing transaction monitoring according to the underground environment transaction monitoring model to obtain a transaction data matrix, and determining underground risk assessment data according to the transaction data matrix; risk quantification is carried out according to the underground risk assessment data, then the transaction risk level of the underground environment is obtained, and abnormity early warning is carried out according to the transaction risk level. According to the method, a more reliable underground environment transaction identification model can be established, and risk quantitative evaluation is carried out based on the dynamic interaction field domain, so that the accuracy of an alarm system is improved.

The application provides a kind of smart terminal virtual sliding button tactile reproduction rendering method and device, belong to touch screen technology and man-machine interaction field. Including the position coordinate of finger on touch screen when user carries out interactive operation, displacement and speed, judge touch click or drag, simulate similar tactile feedback of click, increase tactile braking effect through tactile pawl, according to the size of displacement and movement speed, real-time control the friction coefficient when sliding, produce tactile feedback, realize the tactile reproduction rendering of virtual sliding button. The application produces corresponding tactile feedback in the process of tactile perception through different interactive operations of virtual sliding button, improves the interaction efficiency and accuracy, increases the experience of feeling things with fingertip when using touch screen, reduces the cognitive load of vision and hearing, improves the interaction efficiency and immersion.

The invention discloses a non-contact human-object interaction detection method and system based on a potential interaction field, and the method comprises the steps: inputting a target image to a potential interaction field generator, modeling the spatial context information in the target image into a continuous field, quantifying the potential interaction intensity of an object and a person at any position, and determining the potential interaction intensity of the object and the person at any position; obtaining a potential interactive field mask; inputting the target image and the potential interactive field mask into a feature extractor for feature extraction, and respectively obtaining image features and potential interactive field features; inputting the image features and the potential interactive field features into a fusion encoder for feature fusion to obtain fusion features; semantic understanding is enhanced through a context sensing module, and enhanced features are obtained; and inputting the enhanced features into a decoder for decoding to obtain a human-object interaction detection result. According to the method, the detection capability of the model in a non-contact human-object interaction scene can be improved, and the difficulty of the model in long-distance spatial perception and a multi-human-object pair scene is solved.

The invention discloses a three-electrode eye physiological electricity detection circuit and system. The circuit comprises a first test electrode, a second test electrode, a reference electrode, an acquisition circuit, a signal processing unit and a battery. The first test electrode and the second test electrode are used for reading and transmitting eye physiological electric signals, and the reference electrode is used for denoising the signals; an integrated flexible printed circuit board design is adopted, a differential amplifier, a subtracter, a precise rectifying circuit, an alternating current coupling circuit, a first-order RC low-pass filter and an adjustable resistance amplifier are integrated, and the equipment size is remarkably reduced. According to the invention, a light-weight and high-precision solution is provided for eye physiological electrical detection, the technical blank of home health monitoring is filled, and a new application scene is developed for the man-machine interaction field. The design concept can be further popularized to other physiological electric signal detection systems, and fusion development of medical electronics and intelligent hardware is promoted.

The application discloses a multi-screen linkage VR immersive interaction method and device and medium, and is applied to the VR interaction field. The method comprises the following steps: acquiring three-dimensional scene data of a healing environment; rendering a first visual angle picture based on the three-dimensional scene data, and performing edge splicing on sub-pictures displayed on each display screen corresponding to a multi-screen PC end by using an edge fusion algorithm; establishing a communication connection with a VR end, and sending a healing environment adjustment instruction, so that the VR end adjusts a second visual angle picture displayed according to the healing environment adjustment instruction; or receiving operation state data fed back by the VR end, and checking the operation state data, wherein the healing environment adjustment instruction and the operation state data both comprise corresponding time stamps. It can be seen that the multi-screen and the edge splicing of the sub-pictures displayed on the multi-screen can greatly improve the participation and immersion experience of the same person, and the time stamp mechanism can further realize the synchronization of the visual content of the two ends.

Embodiments of the present application provide a kind of application program operation prompting method, device, computer equipment, computer readable storage medium, computer program product, belong to man-machine interaction field.The method includes: obtaining the moving element track information of application program interface, moving element state information and the key layout information of target keyboard;According to moving element track information and key layout information, establish the key position mapping relationship between the track of moving element and the physical key of target keyboard;Based on moving element state information and key position mapping relationship, determine the target physical key corresponding to target moving element;When target moving element meets prompting trigger condition, output the light prompt signal for target physical key, to control the light emitting element corresponding to target physical key output light prompt.The technical scheme of the present application embodiment can improve the corresponding accuracy between target moving element in application program interface and physical keyboard prompt, and improve the prompt control accuracy for target physical key.

The invention provides a pseudo-materialized man-machine interaction method and device, electronic equipment and a storage medium, and belongs to the field of man-machine interaction, and the method comprises the steps: obtaining the operation state data of each physical equipment in a target power station control system; based on the operation state data, a display interface of a virtual model is driven and updated, the virtual model is constructed according to the actual layout and equipment composition of the target power station control system, and the display strategy of each virtual equipment in the virtual model is determined based on the equipment type or function role of the corresponding physical equipment; in response to a trigger operation for the display interface, generating and issuing a control instruction corresponding to the trigger operation; the control instruction is used for indicating the target power station control system to execute a corresponding control action. The invention can provide a man-machine interaction scheme which is convenient for operators to understand and operate.

This application discloses a biomimetic eye movement control device, method, and robot, belonging to the field of biomimetic robot technology. The device includes multiple flexible electro-actuators, comprising an electro-actuating material that actively deforms under electric field excitation. One end of each flexible electro-actuator is fixed and connected to a support body, while its free end, having a driving displacement or driving force, is directly or indirectly connected to the functional parts of the eye to drive the eyeball shell to rotate and / or drive the eyelids to open and close. A passive constraint structure is also included, configured to provide anisotropic mechanical constraints on the active deformation of the flexible electro-actuators, thereby guiding and converting the active deformation into a driving displacement or driving force along a first preset direction. This solution achieves compact, compliant, low-noise, and highly biomimetic eye coordination control, applicable to humanoid robots, service robots, and human-computer interaction fields.

The application provides a mobile manipulator human-machine safety path planning method based on a danger index; firstly, an external information sensing system of the mobile manipulator is established, including a two-dimensional laser radar system for detecting the distance of surrounding obstacles and a vision system for identifying the human body contour; then, a global map of the working range of the manipulator and the start and end positions are given, and a global path reachable by the mobile manipulator is planned by an algorithm; during the movement process, if the vision system does not detect human body information, the mobile manipulator will move along the global path to the target position; if the vision system detects human body information, the two-dimensional laser radar system is used to scan and obtain the human body contour coordinates, the mobile manipulator and the human body contour are surrounded by plane minimum enclosing circles respectively, the human-machine safety distance is calculated and obtained, and the human-machine danger evaluation index is evaluated; the path planning algorithm is improved, and the mobile manipulator is guided to move away from the personnel in the scene; the application can guarantee the safety of human-machine integration and lay a foundation for the development of the human-machine interaction field.

Technical solutions provide extended reality based capabilities for just-in-time context-based assistance to robotic medical system users. A processor can receive a data stream of a medical session with a robotic medical system. The data stream can be captured by sensors in a medical environment. The processor can determine, based on the data stream and via a context-identification model trained with machine learning, a context of the medical session. The processor can determine, based on the context, an extended reality content to be presented to a user. The processor can detect, based on the data stream and via one or more object-identification models, one or more objects of the medical environment. The processor can establish, based on the objects detected, an interaction field in the medical environment in which to interact with the extended reality content and present the extended reality content to the user for interaction in the interaction field.

The invention discloses an automobile digital cloud exhibition hall interactive experience method based on virtual reality, and relates to the technical field of virtual reality three-dimensional graphic processing, and the method comprises the steps: constructing an interactive automobile digital body, solidifying a movable part motion description record and interactive semantic annotation, and generating an interactive dependence sequence table; obtaining a user experience intention based on the interaction dependency sequence table, converting the user experience intention into a task list, generating a necessary element set for each task, selecting a minimum necessary interaction field from the necessary element set, and generating an interaction script; loading a minimum necessary interaction field according to the interaction script to enter virtual reality, executing actions according to a task list sequence to form candidate vehicle states, calculating a consistency deviation amount, selecting a candidate vehicle state with a smaller consistency deviation amount, and when the script judges that a task is completed, outputting evidence bound with a task type; the reproducible experience record is solidified through the evidence, the interaction script, the action and the vehicle digital body, and script rearrangement is conducted on the task sequence according to the reproducible experience record.

The application discloses a kind of based on virtual reality's car digital cloud exhibition hall interactive experience method, it is related to virtual reality three-dimensional graphics processing technical field, including, constructing interactive vehicle digital body, solidification movable component movement description record and interactive semantic annotation, generate interaction dependent sequence table;Based on interaction dependent sequence table, user experience intention is obtained and is converted into task list, necessary element set is generated for each task, the minimum necessary interaction field is selected from necessary element set, and interaction script is generated;According to interaction script, minimum necessary interaction field is loaded to enter virtual reality, and action is executed according to task list order to form candidate vehicle state, consistency deviation is calculated and candidate vehicle state with smaller consistency deviation is selected, when script determines that task is completed, the evidence bound with task type is output;Through evidence, interaction script, action and vehicle digital body solidification reproducible experience record, and according to reproducible experience record, script is re-arranged to task order.

The application discloses a voice interaction method and device, a vehicle and a storage medium, and belongs to the intelligent interaction field. The method comprises the following steps: receiving a first voice instruction; identifying at least one first node keyword in the first voice instruction; determining a keyword function tree corresponding to each first node keyword from at least one keyword function tree included in a keyword function database based on the at least one first node keyword; in the case that the keyword function trees corresponding to the at least one first node keyword are the same, determining a first target branch node and at least one vehicle-mounted function corresponding to the first target branch node from a first keyword function tree based on the at least one first node keyword; and executing the at least one vehicle-mounted function corresponding to the first target branch node. According to the application, the execution of the vehicle-mounted function can be realized through a single voice, so that the voice interaction efficiency is improved.

The present invention provides a method, device, electronic device and storage medium for action redirection based on dense geometric interaction perception, which relates to the technical field of skinned animation. The method comprises: extracting semantically consistent feature points from the resting postures of a source character and a target character respectively, and calculating the tangent space coordinate system of the feature points to obtain a first feature point grid corresponding to the source character and a second feature point grid corresponding to the target character, performing spatiotemporal representation of the first feature point grid according to the source action sequence of the source character to obtain a source grid interaction field, performing action sequence prediction based on the source grid interaction field, the source action sequence, the first feature point grid and the second feature point grid to obtain a target action sequence of the target character. The method can promote geometric interaction perception action redirection between different grid topologies in a single process, which not only preserves the semantics of motion, but also prevents self-penetration and ensures contact maintenance.

The embodiment of the invention discloses a virtual map display method and device, electronic equipment and a storage medium, and is applied to the field of man-machine interaction, the method comprises the steps that a target interface is displayed, the target interface displays a first virtual map, and the first virtual map displays a virtual role; and in response to a display trigger operation on a second virtual map, displaying at least a partial region of the first virtual map and at least a partial region of the second virtual map on the target interface, the second virtual map displaying a virtual character, and the second virtual map being an independent region entered through the first virtual map. According to the embodiment of the invention, the immersion of the user is improved, the user can understand the moving rule of the virtual map in a more visual and vivid manner, the understanding cost of the user on the target interface in the process of displaying different virtual maps on the target interface is reduced, and the participation experience of the user is optimized while the interestingness is improved.

The invention relates to the technical field of virtual reality and human-computer interaction, in particular to a micro-current tactile parameter automatic generation method and system based on microscopic geometric feature mapping. The method comprises the following steps: acquiring microscopic geometric image data of the surface of a virtual object currently contacted by a user in real time, carrying out explicit feature extraction on the microscopic geometric image data, and calculating to obtain a first visual feature value representing a texture fluctuation change degree and a second visual feature value representing a texture edge sharpness degree; respectively inputting the first visual characteristic value and the second visual characteristic value into a visual tactile roughness mapping model and a flexibility mapping model to respectively obtain a surface space wavelength parameter and a flexibility coefficient of the virtual object; and transmitting the surface space wavelength parameter and the softness coefficient to a micro-current tactile rendering controller, and driving the finger electrode array to generate corresponding tactile feedback. According to the method, the micro-current stimulation control parameters are automatically generated by analyzing the visual features of the virtual textures, and the problems that in the prior art, dependency on a pre-stored database is high, and adaptability to unknown textures is poor are solved.

The invention discloses a false touch prevention interaction method based on medium-free holography, and particularly relates to the field of false touch prevention interaction of medium-free holography, which comprises the following steps: generating a medium-free holographic image through a medium-free holographic imaging device, the semantic interaction body is composed of a plurality of interaction voxels or interaction planes; acquiring a human face image through a camera arranged on the medium-free holographic imaging equipment, and identifying a human face position, a human face orientation, a human face size and an eye distance parameter based on the human face image so as to determine a spatial range of a human face in the medium-free holographic image; when the face is in the spatial range, determining a first duration when the face is in the spatial range; an operation pre-confirmation mechanism is introduced into a multi-level judgment chain of face recognition, sight tracking, space-time intention judgment and mode comparison, so that the real intention of a user is verified before interaction triggering.

The application discloses a sharing method and device of information flow, equipment and storage medium, belong to man-machine interaction field. Method includes: obtaining the clipping information sent by the first client, the first client includes the reading function of information content from information source, the clipping information is the relevant information provided by the first user account when customizing clipping; at least one information flow element is generated according to the clipping information, and the information flow element includes information content and information source;According to at least one information flow element, the information flow of the custom clipping corresponding to the first user account is generated;In response to receiving the sharing request of the first client, the information flow of the custom clipping is sent to at least one second client. Through the generation of the information flow of the custom clipping, the information of multiple sources can be collected, and the efficiency of information collection is improved. At the same time, the information flow is stored in the server, so that the client can quickly read the target information flow according to the index, and the efficiency of file management is improved.