10results about How to "Minimize overhead" patented technology

The invention provides a configurable caching method and system for an artificial intelligence reasoning process. According to the method, a cache interface layer is introduced to be integrated with an artificial intelligence framework, flexible cache strategies including a cache key generation method, a storage mode, cache capacity, an elimination strategy and the like are defined by utilizing a user configuration module, and a cache key generator is used for generating cache keys according to the strategies. The cache storage manager is responsible for managing storage, retrieval and updating of cache data and supports various storage rear ends. In the reasoning process, the cache querying and updating module is responsible for querying and updating the cache. The system can also comprise a performance monitor and a data security module which are respectively used for monitoring cache performance indexes and ensuring data security. The corresponding system comprises a cache interface layer, a user configuration module, a cache key generator, a cache storage manager and a cache query and update module. According to the method, the artificial intelligence reasoning efficiency can be improved, the resource utilization is optimized, a flexible and configurable cache solution is provided, and the method is easy to integrate with a mainstream artificial intelligence framework and adapts to different task requirements.

The invention discloses a road vehicle detection method and system based on improved YOLOv7. The method comprises the following steps: step 1, collecting vehicle image data on a road; step 2, preprocessing the collected vehicle image data; 3, making a vehicle data set after preprocessing, and dividing the vehicle data set into a training set and a verification set; 4, training the data set by using a YOLOv7 model to obtain a vehicle target detector; and step 5, using the vehicle target detector obtained in the step 4 to detect a to-be-tested image or video to obtain a detection result. According to the method, the problems of low detection speed, low detection precision, complex background interference and the like in a road vehicle detection task are solved, and the requirements in an actual scene can be well met.

The invention discloses a target tracking-oriented multi-base radar receiving station distribution and networking architecture optimization method. The method comprises the following steps of: firstly, acquiring a hierarchical distributed Cramer-Rao lower bound considering bilateral packet loss; secondly, combining the radar node resource overhead model and the communication resource overhead model to obtain a multi-base radar resource overhead model; and then according to the hierarchical distributed Cramer-Rao lower bound, establishing a target tracking precision model with resource variables. And finally, based on the multi-base radar resource overhead model and the target tracking precision model, a multi-base radar receiving station allocation and networking architecture optimization model is established and solved, and according to a solving result, multi-base radar receiving station allocation and networking architecture optimization is completed. According to the method, the radar network resource overhead is minimized under the condition that the preset multi-target tracking precision is ensured, and the method can be well applied to the field of target tracking and information fusion.

This invention relates to a preemptive scheduling method for heterogeneous multi-core processors based on multi-agent reinforcement learning. It includes: constructing a system model comprising a set of sporadic real-time tasks and a heterogeneous multi-core platform; pre-setting fixed preemption points for tasks; dividing task execution into non-preemptive segments; modeling the scheduling problem as a partially observable Markov decision process; designing state representations that include task dynamic attributes and platform heterogeneity characteristics; wherein the task dynamic attributes are normalized values ​​of relative deadline pressure, execution progress ratio, and relaxation; and the platform heterogeneity characteristics include a weighted sum of core computing power level and global idle computing power; clustering processors according to computing power; assigning independent agents to each cluster; and adopting a fully cooperative mode to share global rewards; employing a centralized training and distributed execution framework; using the discrete soft actor-critic algorithm during training; with each agent's policy network collaboratively optimizing under the guidance of a centralized value function; and during execution, each agent independently deciding on preemption timing based on local observations. This invention achieves cooperative preemptive scheduling among heterogeneous cores through multi-agent reinforcement learning, effectively reducing task overdue rates and system overhead, and improving real-time performance.

A video signal processing method and apparatus for encoding or decoding a video signal are disclosed. More specifically, the present invention provides a video signal processing method and a video signal processing apparatus using the same, the method comprising: receiving intra-frame prediction mode information for a current block, the intra-frame prediction mode information indicating one of a plurality of intra-frame prediction modes constituting an intra-frame prediction mode set; and decoding the current block based on the received intra-frame prediction mode information, wherein: the intra-frame prediction mode set includes a plurality of angle modes; the plurality of angle modes include a basic angle mode and an extended angle mode; and the extended angle mode is sent with a signal based on the basic angle mode.

A user equipment and a method performed by the user equipment are disclosed. The method includes the following steps: receiving system information for configuring a first K offset as a scheduling offset for correcting the timing of uplink signals for NTN; receiving a control signal for configuring or updating a second K offset based on the first K offset; and transmitting uplink signals based on the first K offset or the second K offset, wherein the uplink signals are transmitted based on the first K offset in association with a random access channel (RACH).

A method for communicating with a macrocell base station and a small cell base station through a user equipment in a communication system is provided, and the method includes the following steps: applying a first base station security key to a first communication link with the macrocell base station; generating a second base station security key to be used for a second communication link with the small cell base station based on the first base station security key; applying the second base station security key to the second communication link with the small cell base station; and communicating through at least one of the first communication link and the second communication link.

Disclosed is a method for supporting adaptive unloading of multi-Internet of Things (IoT) applications in an edge environment, including the following steps: constructing an application compliant with a universal program structure supporting on-demand computing unloading; for the application compliant with the universal program structure supporting on-demand computing unloading, extracting a program fragment flowchart through static code analysis to provide internal flow information of the application for generation of an unloading solution; analyzing a peripheral mobile edge computing (MEC) environment and the application through an unloading solution generating algorithm based on a multi-task particle swarm optimization-genetic algorithm to obtain the optimal unloading solution; and performing, by the application, computing unloading by taking a service as a particle size according to the unloading solution to minimize a system overhead under a circumstance that a deadline constraint of each application is satisfied. The method supports the computing unloading of different types of applications in the MEC environment, and minimizes the system overhead generated by the computing unloading.

The utility model relates to the technical field of silicon carbide devices, and discloses a silicon carbide semiconductor device integrated with a temperature sensor, which comprises an N-type silicon carbide substrate layer (102) and an N-type silicon carbide epitaxial layer (103), at least two P + injection regions (105) are arranged on the N-type silicon carbide epitaxial layer (103), N + injection regions (109) are arranged on the P + injection regions (105), a source electrode conducting layer (106) is arranged on the P + injection region (105) on one side, and a source electrode conducting layer (106) is arranged on the N + injection region (105) on the other side. And a T-pole conductive region (110) is arranged on the P + injection region (105) on the other side. According to the utility model, the temperature sensor is integrated into the silicon carbide device, the diode is used as the temperature sensor and is integrated into the source region of the silicon carbide device, and the T electrode is independently added for the sensor, so that the high-temperature-resistant temperature sensor is realized, and the Vf value and the temperature of the diode have good linearity; in addition, the influence of the sensor on other parameters of the silicon carbide device is eliminated, the layout overhead is minimized, and good compatibility is achieved.

An adaptive link management method and system based on satellite mobile communication employs a PPP-based UDP protocol for initial link negotiation. Network address and other information are exchanged through link establishment request and confirmation signaling interactions. Network segment relationships are determined based on address information: GRE encapsulation with proxy ARP assistance is used for same network segments, while standard GRE encapsulation is used for different network segments. Link endpoint identifiers are negotiated and a bidirectional transmission link is established under the selected encapsulation mode. After link establishment, both parties periodically exchange keep-alive signaling to maintain connectivity. Simultaneously, the status of connected local access devices is monitored, and information is synchronized according to a preset period and / or status changes to dynamically update the transmission path and perform IP address conflict detection and handling. The transmission link is released normally through link release signaling. This scheme improves link adaptability and bandwidth utilization, effectively prevents address conflicts, and thus enhances network reliability.