THz Blockage Prediction Using LSTM and Beamforming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human blockage models for THz communications, such as the double-truncated multiple knife-edge (DTMKE) diffraction model, are inadequate for predicting blockages in highly directional THz channels due to their reliance on omnidirectional transmission and reception assumptions, which become inaccurate with beamforming.
Innovation Solution
A method and apparatus using Long-Short-Term Memory (LSTM) to predict channel state and potential blockages by simulating a human body blocker, determining multiple stages of potential blockage, and deciding on conditional handovers based on the predicted channel state, incorporating both DTMKE model and beamforming equations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional human blockage models (DKED, MKED, DTMKE) are used for THz communications, then the models provide a framework for blockage analysis, but they become inaccurate when applied to highly directional THz channels with beamforming due to their omnidirectional transmission assumptions
Solution Approach 1:
The patent modifies the conventional DTMKE model by incorporating beamforming parameters specific to THz communications. The model transitions from omnidirectional transmission assumptions to directional beam patterns by introducing beamwidth, beam direction, and antenna gain parameters that characterize highly directional THz channels. This parameter adaptation enables the model to accurately predict blockages in beamformed systems while maintaining the geometric diffraction framework.
2Loss of time
If LSTM prediction is implemented for blockage detection, then early blockage prediction capability is achieved, but computational complexity and processing requirements increase
Solution Approach 1:
The patent implements LSTM-based prediction to detect potential blockages before they completely occlude the LoS path. By training the LSTM model on historical channel state information and blocker trajectory data, the system learns to predict future blockage events ahead of time. This preliminary detection enables proactive handover decisions, allowing the system to switch to alternative paths or adjust beam directions before link failure occurs, thereby reducing effective loss time despite increased computational requirements.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy of blockage prediction and handover decisions in THz communications, improving system performance by accounting for the actual shape of the human body and directional beam patterns, thus reducing link failures and enhancing user equipment localization.
Implementation Method 1
predicting, using LTSM, a channel state and the potential blockage
Implementation Method 2
a double-truncated multiple knife edge (DTMKE) diffraction model is considered
Data Source
AI summary
Disclosed is a method for determining whether to trigger of a conditional handover, including estimating, using a line-of-sight (LoS) channel, a received signal power as a function of a blocker used to simulate a human body, determining multiple stages of potential blockage of the received signal power by the blocker, predicting, using long-short-term memory (LTSM), a channel state and the potential blockage, and determining whether to trigger the conditional handover based on the predicted channel state and potential blockage.