7results about How to "Reduce bandwidth utilization" patented technology

The present disclosure provides an analytic personalization framework. Methods, systems, and computer readable media for personalized analysis of user interfaces. The method includes generating a training dataset from received user interaction data, inputting the training dataset to a machine learning model to train the model, generating a set of user interest scores for a particular user, where each user interest score indicates a user's interest to access information corresponding to a UI element of an application, and generating a user interest score for the particular user. The method includes determining a user interest score, determining that the user is interested in a particular UI element that is not included in the initial UI and has at least a threshold score according to the user interest score, dynamically modifying the initial UI to include the particular UI element, presenting an updated UI, monitoring further user interactions, updating a model based on the further user interactions, and modifying the updated UI based on the updated model.

This invention relates to the field of Internet of Things (IoT) technology, providing a large-scale IoT system, method, and medium for remote firmware updates of sensors. The system includes a sensor management platform configured to: determine multiple sensor update groups based on data characteristics of multiple sensors and their geographical locations within an industrial grid area; determine the update priority of each sensor update group based on its task and transmission characteristics; determine update parameters for the multiple sensor update groups based on update priority, firmware characteristics, and channel state distribution; and generate update instructions and firmware update packages based on the update parameters, and control multiple sensors in the multiple sensor update groups to perform firmware updates. The method can be executed after being read from computer instructions stored on a computer-readable storage medium. This invention enables secure, reliable, and continuous large-scale sensor firmware upgrades under complex industrial network conditions.

The invention discloses a vehicle-mounted gateway cross-bus signal splitting and selective forwarding method and system, and relates to the technical field of vehicle-mounted network communication, and the method comprises the steps that a gateway ECU receives a network signal message and analyzes the network signal message to obtain an analyzed network signal, and the network signal is obtained from a CAN network or a CAN-XL network; selecting and splitting the analyzed network signal, selecting a network signal needing to be forwarded, and filtering and splitting the network signal not needing to be forwarded to obtain a processed network signal; setting a plurality of priority gradients based on the network signal message, generating a plurality of queues based on the plurality of priority gradients, and inputting the processed network signal into the queues; and setting a time-based trigger condition, packaging the signals in the queue into a new CAN message or a CAN-XL message after the signals in the queue reach the time-based trigger condition, sending the new CAN message or the CAN-XL message to the gateway ECU, and emptying the queue, thereby realizing signal splitting and selective forwarding.

Systems, devices, and methods for implementing memory bandwidth reduction techniques for low-power convolutional neural network inference applications are disclosed. A system includes at least a processing unit and an external memory connected to the processing unit. The system detects a request to perform a convolution operation on input data from multiple channels. In response to detecting the request, the system partitions the input data from the multiple channels into 3D blocks so as to minimize external memory bandwidth utilization for performing the convolution operation. Next, the system loads the selected 3D blocks from the external memory into an internal memory, and then generates convolution output data for one or more features for the selected 3D blocks. Then, for each feature, the system adds the convolution output data together across channels before writing the convolution output data to the external memory.

The present invention discloses a credit value optimization method for a credit-based shaper in a time-sensitive network, which belongs to the field of communications and includes the following steps: real-time monitoring of time-sensitive network links to obtain link load, frame waiting time and traffic priority information; if the link load reaches a preset threshold, dynamically adjusting the recovery rate of the CBS credit value, and accelerating the recovery of the credit value by adjusting the idleSlope; using a heuristic algorithm, combining the dynamic changes of the link load and the frame waiting time, and adaptively optimizing the idleSlope; performing priority management on the AVB stream to ensure that high-priority traffic is transmitted first, and reducing the waiting time of the AVB stream by accelerating the recovery of the credit value. The present invention introduces parameters such as frame waiting time on the basis of logical bandwidth allocation, and the new credit value recovery mechanism significantly reduces the AVB stream transmission delay and increases bandwidth utilization.

This application is applicable to the field of communication control, and provides a power line communication method, a primary routing device, and an apparatus. In the method provided in this application, the primary routing device may obtain a device access status of a first secondary routing device. If the device access status of the first secondary routing device is a first state, it indicates that no user equipment accesses the first secondary routing device. In this case, the primary routing device may allocate a transmission opportunity not in a first time segment but in a second time segment of a transmission management cycle to the first secondary routing device. This reduces transmission opportunities of the first secondary routing device, avoids a waste of excessive time slots by the first secondary routing device, improves bandwidth utilization of a power line network, and has strong practicability and usability.

The present invention discloses a target detection method and system based on a cloud-edge collaborative hybrid architecture, which belongs to the field of cloud-edge collaborative technology. The technical problem to be solved is how to achieve target detection while ensuring edge security in edge scenarios with limited computing power or hardware. It includes a global manager and a cloud-side worker component deployed on a cloud node, and an edge controller and an edge-side worker component deployed on an edge node; the cloud-side worker component is used to perform cloud-side model training on the target detection model; the edge-side worker component is used to obtain samples from edge devices based on AI tasks, and is used to perform edge-side reasoning calculations through the target detection model specified by the AI ​​task, and display edge-side reasoning results; the cloud-side worker component is used to perform cloud-side reasoning on difficult samples, and the edge-side worker component is used to display cloud-side reasoning results.