6 results about "Alternating least squares" patented technology

The application discloses a deep learning driven target parameter estimation method, and relates to the cross field of radar signal processing and artificial intelligence. The application embeds the maximum correlation entropy criterion into an alternating least square process, expands the alternating least square process into a differentiable network layer containing weight updating and a damped weighted least square, realizes end-to-end optimization under a unified criterion by combining a deep initialization network, introduces a sigma-Net based on a teacher-student framework, dynamically predicts an optimal kernel width based on three types of characteristic features, namely, a residual tail ratio, a peak ratio and an estimated GSNR, and synchronously optimizes initialization quality and kernel width strategies through double-path gradient back propagation, so that the synergistic gain of high-quality initialization, robust updating and adaptive parameter adjustment is realized.

The present application relates to resource recommendation technical field, especially to a kind of learning resource recommendation method and system based on user portrait, including establishing user portrait matrix and learning resource feature matrix, and using factor decomposition machine and collaborative filtering algorithm to build the implicit semantic matching model of user and learning resource;The parameter of implicit semantic matching model is learned by alternating least squares method, and user-learning resource implicit semantic association feature is generated;User-learning resource implicit semantic association feature is used as individual code, and the Pareto optimal recommendation solution set of multi-objective learning resource recommendation optimization model is searched using non-dominated sorting genetic algorithm, and learning resource recommendation scheme is generated in combination with optimal recommendation decision model.The present application dynamically adjusts recommendation strategy by combining multi-objective optimization and deep reinforcement learning etc., improves the accuracy of recommendation, while it can balance the diversity and personalization of recommended content, improve the dynamic adaptability and robustness of recommendation system.

This invention relates to a method for simultaneous localization and mapping (SLAM) assisted by a UAV equipped with a reconfigurable smart surface (UAV-RIS), addressing the problems of parameter decoupling difficulties caused by cascaded channels and large amplitude-based localization and mapping errors due to signal strength distortion. The method includes: 1) constructing a nested parallel factor model of the received signal and performing tensor decomposition using a bilinear alternating least squares method to achieve joint channel parameter estimation and signal detection; 2) initializing and removing outliers using an amplitude-independent geometric inference method based on the estimated geometric parameters; and 3) optimizing the model using a weighted least squares method to determine user state and environmental characteristic parameters. This method reduces the impact of amplitude model mismatch and improves parameter estimation accuracy in complex channel environments, making it suitable for autonomous localization and mapping in low-altitude wireless networks.

The application discloses a kind of intelligent reflecting surface auxiliary cascading surface array channel nested tensor decomposition estimation methods, first constructs cascading channel model, initialization each end guiding matrix and channel matrix in turn;Zero intelligent reflecting surface phase shift is first placed, respectively constructs the tensor model of transceiver end parallel factor decomposition, combined with three linear alternating least square method and least square method, completes two-dimensional angle estimation of transceiver end;Again control phase shift, through two linear alternating least square method, Khatri-Rao decomposition, complete intelligent reflecting surface two-way angle estimation, finally rely on Kronecker product and Khatri-Rao product model solves beam gain.The application relies on nested parallel factor decomposition, realizes four angles automatic pairing, adapts single, multi-beam scene, reduces high-dimensional calculation amount, suppresses off-grid energy leakage, without signal autocorrelation matrix and complex angle pairing, adapts super large scale surface array, improves channel estimation applicability, supports sensing integration system design.

This invention provides a deep denoising-based multi-RIS indoor positioning method. For indoor scenarios with multipath interference and noise, this invention first constructs a fourth-order tensor model from the channel state information reflected from multiple indoor RIS points and scattering points to the receiver. Second, it uses antenna rearrangement technology to rearrange the tensor model to satisfy the decomposition uniqueness condition. Then, it utilizes the denoising capabilities of a deep learning architecture to suppress noise in the received signal. Further, it employs an optimized quadlinear alternating least squares algorithm to estimate the factor matrix and extract channel parameters. Finally, it uses a geometrically based search-free spatial positioning method to estimate the user's location and azimuth angle, and map the indoor environment. Therefore, the deep denoising-based multi-RIS indoor positioning method proposed in this invention has higher accuracy and robustness compared to comparative algorithms, and is more in line with the needs of practical communication scenarios.

This invention provides a terahertz massive MIMO wireless simultaneous localization and mapping method to address the difficulties in decoupling multidimensional parameters due to the dual-wideband effect and the data association ambiguity caused by the lack of Doppler information in dynamic scenarios. The implementation steps are as follows: 1) Construct a third-order tensor model of the received signal containing spatial, frequency, and time dimensions, and expand the tensor using a time smoothing strategy to overcome the rank deficiency problem; 2) Iteratively decompose the observation tensor in the compressed domain using a trilinear alternating least squares algorithm to achieve high-precision channel parameter decoupling and communication signal detection; 3) Encapsulate the estimated Doppler, angle, and time delay into measurement vectors, calculate the joint posterior probability based on the confidence propagation algorithm, eliminate data association ambiguity, and infer the trajectory of the mobile agent and environmental characteristics. This invention effectively overcomes the dual-wideband effect, achieving accurate tracking and robust mapping.