19 results about "Ubiquitous computing" patented technology

The invention provides a multi-user computing resource allocation method and device and a medium. Comprising the following steps: acquiring scene information of a to-be-distributed multi-user low-orbit satellite ubiquitous calculation scene; determining an optimization objective function and an overall optimization problem according to the scene information; establishing a first agent used for access selection and a second agent used for calculation unloading; determining a first reward function of the first agent and a second reward function of the second agent according to the optimization objective function and the overall optimization problem; performing centralized training on the first intelligent agent and the second intelligent agent by adopting an asynchronous alternating algorithm to obtain a third intelligent agent parameter corresponding to the first intelligent agent and a fourth intelligent agent parameter corresponding to the second intelligent agent; according to the third agent parameter and the fourth agent parameter, determining an access selection decision and a calculation unloading decision of a to-be-allocated multi-user low-orbit satellite ubiquitous calculation scene, and performing calculation resource allocation according to the access selection decision and the calculation unloading decision; therefore, the resource allocation efficiency is improved, and the time delay and the energy consumption are reduced.

The present application relates to devices and components including apparatus, systems, and methods for ubiquitous computing.

The present invention provides an edge-based generative large language model collaborative reasoning method for a ubiquitous computing power network, comprising obtaining generative large language model information and edge device information; modeling and solving edge device pipeline orchestration and generative large language model pipeline parallel reasoning; pipeline parallel pre-filling based on input sequence division; and pipeline parallel autoregressive decoding based on an outline. The present invention, with the help of computing resources of multiple edge devices in a ubiquitous computing power network environment, accelerates the pre-filling stage of the generative large language model through a pipeline parallel reasoning technology based on sequence division, while utilizing an outline-based pipeline parallel decoding method to accelerate the autoregressive decoding stage of the generative large language model, thereby achieving low-latency and resource-efficient distributed collaborative reasoning of the generative large language model on multiple edge devices; effectively protecting user privacy data security, the present invention has better adaptability and robustness for the resource characteristics of complex edge environments.

The application belongs to the technical field of ubiquitous computing. The application provides an end-side hyperspectral imaging and fine identification method for an open environment. The disclosure embodiment designs a spatial spectrum two-type modal encoder for extracting spatial spectrum features of RGB and NIR images and outputting a fused feature stream; a light hyperspectral imaging model is used to process the fused feature stream and reconstruct a hyperspectral image. A light decoupling loss function design is used to replace the original loss function of the light hyperspectral imaging model, and robust hyperspectral image imaging in an open environment is realized. An analysis module based on grouping attention is designed, different weights are assigned to different absorption peaks of different substances by grouping along the spectral dimension, and the huge calculation amount and calculation delay caused by full-band analysis are avoided. A mixed quantization method is used to reduce the calculation amount and storage amount of the model deployed on the end side without obvious decrease in imaging and analysis accuracy of the model.

The application provides an artificial intelligence platform resource metering management method, specifically comprising the following steps: S1: detecting and metering the computing power surplus of each computing power point in the platform system, and dividing the computing power points into training points and reasoning points according to the idle computing power surplus and the computing power point type; S2: detecting and metering the storage performance indicators of each storage point in the platform system to obtain the parameters of each storage point; S3: evaluating the algorithm model in the platform system; S4: detecting and evaluating the business data in the platform system; S5: selecting the computing power points, storage points, algorithm model and business data to form a task combination, so that the utilization rate and benefit of various resources are optimal. In the application scheme, the computing power and network resources are integrated into a whole to form an efficient, scalable and flexible ubiquitous computing power network, the cluster advantage of the computing power is exerted, the scale efficiency of the computing power is improved, and the global intelligent scheduling and optimization of the computing network resources are realized.

The anti-environmental noise tooth gesture interaction action independent identity authentication method belongs to the field of Internet of Things and ubiquitous computing. The ear outside and ear inside microphones of the intelligent earphone respectively perceive environmental noise signals and ear canal audio signals. The ear inside collected audio signals have frequency selection and individual difference characteristics due to the effect of ear closing plug. When the user performs tooth gesture interaction action, the corresponding ear inside bone conduction audio signals exist individual differences after modulation through the facial transmission path, and are not easily affected by the interaction action. The present application effectively eliminates environmental noise interference by constructing individualized mapping relationship of ear inside and outside audio signals, and uses the adversarial learning network to extract the transmission path tissue related biological features from the captured tooth gesture ear inside bone conduction audio signals, realizes the anti-environmental noise tooth gesture interaction action independent identity authentication. The present application has the advantages of fast authentication, effective resistance to environmental noise and robustness to inconsistent tooth gesture interaction action.

The invention provides a cooperative implementation method for vehicle tasks and a vehicle, and the method comprises the steps: obtaining equipment parameters of terminal equipment in the vehicle, and determining a trusted equipment list according to the equipment parameters, and the trusted equipment list at least comprises one piece of trusted equipment; obtaining task information of a target task of the vehicle, and dividing the target task into a plurality of sub-tasks based on the trusted device list and the task information; and selecting a target trusted device corresponding to each sub-task from the trusted device list according to the task information, and allocating the sub-tasks to the corresponding target trusted devices. According to the method, the reliability, efficiency and quality of vehicle task collaboration can be improved by constructing the in-vehicle near-field low-delay self-organizing ubiquitous computing power network.

Embodiments of the present application provide a computing power scheduling method and device and network equipment. The method comprises: obtaining computing power update information sent by at least one network node in a computing power network system; and performing computing power resource allocation according to the computing power update information. By using the method, computing power update information sent by network nodes at all levels in the computing power network system can be obtained in real time, user demand can be analyzed according to the computing power update information, and dynamic computing power resource allocation can be realized, so as to avoid the problems of computing power island of edge computing power nodes and large number of ubiquitous computing devices that cannot be fully utilized.

The invention discloses an anti-environmental noise tooth gesture interaction action irrelevant identity authentication method, and belongs to the field of Internet of Things and pervasive computing. An extraauricular microphone and an intraauricular microphone of the smart earphone sense an environmental noise signal and an audio signal in an ear canal, respectively. The related occlusion effect enables audio signals acquired in the ear to have frequency selection and individual difference characteristics. When the user executes the tooth gesture interaction action, the corresponding in-ear bone conduction audio signal has individual difference after being modulated by the face conduction path, and is not easily influenced by the interaction action. According to the method, the environmental noise interference is effectively eliminated by constructing the personalized mapping relation of the audio signals inside and outside the ear, and the confrontation learning network is utilized to extract conduction path tissue related biological characteristics from the captured tooth gesture in-ear bone conduction audio signals, so that the tooth gesture interaction action irrelevant identity authentication resisting the environmental noise interference is realized. The method has the advantages of fast authentication, effective resistance to environmental noise and robustness to inconsistent interaction actions of tooth gestures.

The application discloses a ubiquitous computing power sharing method and device, equipment and medium, wherein the ubiquitous computing power sharing method detects and configures the system of a target device through the built-in operating system of a non-intrusive ubiquitous computing power sharing hardware medium, uploads the hardware resource information of the device to a cloud platform, virtualizes idle disk sectors into virtual disks, and synchronizes the resource information of the virtual disks to the cloud; the cloud platform analyzes the resource condition of the target device, and allocates appropriate computing power tasks to the device; the built-in operating system creates multiple virtual machines on the target device and executes the computing power tasks, and finally returns the task result information to the cloud platform; after the cloud platform verifies that the result is valid, the built-in operating system automatically restores the original state of the target device. Efficient and safe computing power sharing of the device that can be connected to the non-intrusive ubiquitous computing power sharing hardware medium is realized.

An apparatus configured to generate, for transmission to a network, a request for computing resource availability for a computing task to be offloaded from the apparatus, wherein the request comprises an indication of a Ubiquitous Computing Characteristics Profile (UCCP) comprising one or more attributes related to the computing task and process, based on signals received from the network, computing node information associated with a network node, wherein the apparatus offloads the computing task to the network node.

The application discloses a music playing mode intelligent switching system suitable for different scenes and belongs to the field of artificial intelligence and ubiquitous computing. The system comprises a multi-modal perception module, which is used for collecting environment, user physiological, behavior interaction and audio content data in real time; a feature fusion coding module, which generates a joint scene feature vector through time sequence alignment, statistical feature extraction and neural network coding; a machine learning inference engine, which adopts a pre-trained multi-task learning model supporting online fine-tuning, outputs a fine-grained scene recognition result and optimal playing mode configuration parameters according to the feature vector; and a dynamic audio execution module, which analyzes the configuration parameters and realizes non-perception audio experience switching through a smooth transition controller, and simultaneously collects user feedback to form a learning closed loop. The application realizes highly personalized, context-accurate perception and smooth and seamless music playing mode full-automatic intelligent adaptation, and can be widely applied to intelligent earphones, vehicle entertainment, smart home and mobile terminals.

Embodiments of the present invention provide a knowledge information analysis method and system based on ubiquitous computing. For online behavior extracted knowledge information, based on a single traversal loop, a feature extraction unit in a model unit of a penalty-based feature extraction model generates at least one candidate behavior extraction feature corresponding to a third ubiquitous computing target based on the behavior extraction features of two of the three ubiquitous computing targets: a first behavior entity, a second behavior entity, and an entity relationship in the online behavior extracted knowledge information. This determines a candidate behavior extraction feature, and ubiquitously computes the online behavior extracted knowledge information based on the obtained target behavior extraction features of each behavior entity or entity relationship in the online behavior knowledge network. This allows for effective ubiquitous computing of online behavior extracted knowledge information, while highlighting entity relationships at different knowledge levels.

The invention provides a remote control method and system for ubiquitous hardware equipment. The remote control method comprises the steps that connection between a client side and a remote server side is established; the method comprises the following steps of: capturing an I / O (Input / Output) request of Virtio virtual equipment at a client, serializing and packaging the I / O request to obtain a Virtio request network data packet, and sending the Virtio request network data packet to a remote server; at the remote server side, receiving the Virtio request network data packet, performing analysis and deserialization processing on the Virtio request network data packet, processing an obtained original I / O request, serializing a processing result, packaging the processing result into an I / O request network data packet, and returning the I / O request network data packet to the client side; and the client receives the I / O request network data packet, analyzes and deserializes the I / O request network data packet, and finally transmits a processing result to a corresponding application program. According to the method, the management efficiency of heterogeneous hardware resources in the ubiquitous computing environment and the flexibility of system resources are improved.

The present application relates to devices and components including apparatus, systems, and methods for ubiquitous computing.

This invention discloses a method for scheduling computing resources in ubiquitous computing networks (UCPN). Starting from the topology and resource characteristics of UCPN networks, a unified UCPN hypergraph model is proposed. This model consists of heterogeneous nodes with differentiated resources and hyperedges formed by inter-node dependencies and supply capabilities. Each hyperedge contains a set of heterogeneous nodes with similar resource supply capabilities, rather than a bilateral relationship between two individuals, thus improving node selection and resource mapping speed. Based on the constructed UCPN hypergraph model, a hypergraph convolutional attention network for UCPN computing resource scheduling is proposed. This network achieves end-to-end data transmission path planning while scheduling computing nodes through hypergraph convolution and graph attention mechanisms. Simulation results show that the proposed HGcov-atten outperforms benchmark schemes in terms of task execution time and resource utilization.

The invention provides a palm virtual touchpad method and device based on finger positioning and tracking. According to the invention, the measurement of contact information such as absolute positioning, touch detection and continuous pressing force of a finger touching on a palm is supported; meanwhile, the problems that in the ubiquitous pervasive computing environment, a traditional input device is not portable, nowhere to place, difficult to use at any time and the like are solved, the palm of a user serves as an input medium, calibration and complex sensor deployment of the user are not needed, system setting is remarkably simplified, and the user experience is improved. Compared with the prior art, the method has the advantages that interaction flexibility and accessibility are improved, good universality and expandability are achieved, the method is particularly suitable for input requirements in mobile and wearable devices or other limited interaction spaces, and a natural, efficient and always available man-machine interaction solution is provided for a complex and changeable pervasive computing environment.

A distributed computing management function (DCMF) is configured to receive computing resource availability information from a first network node and transmit computing node information associated with the first network node to a second network node, wherein the second network node offloads one or more computing tasks to the first network node and the data for the one or more computing tasks does not traverse a core network.