5 results about "Temporal multiplexing" patented technology

A kind of multi-path serial exchange FFT IP core based on time multiplexing constant multiplier belongs to digital signal processing technical field, solve existing time multiplexing constant multiplier when adapting MSC architecture FFT, there is the problem of collaborative conflict with serial exchange data scheduling, hardware resource redundancy, poor point number expansibility, timing and signal quality imbalance and lack of supporting efficient data permutation scheme, including multi-path serial exchange FFT architecture, the architecture includes: memory-based permutation module, a plurality of general rotator and at least two base modules, wherein k is positive integer;The rotator in the base module is optimized using time multiplexing constant multiplier, and the time multiplexing constant multiplier is used to implement the multiplication operation of fixed constant and variable input;The IP core supports the FFT calculation of point number 2 power exponent.This application is suitable for digital signal processing scene.

This invention relates to the field of wireless communication and target detection technology, and discloses a spatiotemporal multiplexing adaptive beam coordination control method and system for a RIS-assisted ISAC system. The method acquires channel state, user location, target azimuth, obstruction flags, multipath parameters, RIS hardware constraints, and task priority information to construct a parameterized model; allocates communication and sensing time slots in the time domain and divides communication and sensing functional subarrays in the spatial domain; establishes a joint optimization objective, employs hierarchical robust alternating optimization to obtain the continuous-domain RIS configuration, and maps it to a discrete codebook for robustness verification; adaptively updates the RIS configuration using a bidirectional closed-loop feedback and a three-level triggering mechanism, outputting the RIS configuration matrix, spatiotemporal resource allocation scheme, and beam coordination control parameters. This invention can improve the communication-sensing coordination capability, system robustness, and burst response capability under complex obstruction, multipath disturbances, and hardware non-ideal scenarios.

The application discloses a quality-of-service oriented GPU space-time adaptive resource scheduling method and system, belongs to the field of heterogeneous resource scheduling and management under deep learning inference, and comprises the following steps: determining a minimum SM subset satisfying a quality-of-service constraint and corresponding SM mask configuration in an offline stage; collecting multi-task co-location operation data offline, and constructing an interference-aware delay prediction model by using the multi-task co-location operation data; in an online scheduling stage, when a new task arrives, it is judged whether the GPU is in an idle state according to a current GPU load state; the constructed interference-aware delay prediction model is used for prediction, and it is judged whether the predicted inference delay and the predicted delay both satisfy preset quality-of-service constraints; a time multiplexing strategy is adopted to schedule and execute tasks by intercepting a kernel-level command buffer, and it is ensured that only one task occupies all resources of the GPU at the same time. The application can improve system throughput and guarantee the quality of service of delay-sensitive tasks.

The application relates to the technical field of image analysis, and particularly relates to an image analysis device based on deep learning. The device comprises a model warehouse, a running environment monitor, a model decision engine, a heterogeneous computing engine and a unified inference interface; the model warehouse stores a plurality of deep learning models for the same image analysis task; the running environment monitor is arranged on a target device and is used for collecting system state indexes of the target device in real time; the model decision engine is coupled with the model warehouse and the running environment monitor respectively and is used for dynamically generating control instructions based on a model predictive control algorithm during task execution; the heterogeneous computing engine is used for receiving the control instructions, loading corresponding models from the model warehouse according to model version identifiers in the control instructions, and executing an image analysis task according to the space allocation intensity coefficient and the time multiplexing intensity coefficient.

This invention provides a radar device, a radar signal generating device, and a radar signal generating method that can improve the detection accuracy of targets. The radar transmitting unit (100) includes: multiple transmitting antennas for transmitting signals; and a transmitting circuit that multiplexes the transmitting signals from the multiple transmitting antennas by imparting a phase rotation amount corresponding to a Doppler frequency shift to the transmitting signals. The transmitting circuit performs code multiplexing of the transmitting signals corresponding to a first Doppler frequency shift and uses different transmitting antennas to perform time multiplexing of the transmitting signals corresponding to a second Doppler frequency shift that is different from the first Doppler frequency shift.