Neural Network Preamble Detection for Wireless Access Nodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in accurately detecting random access preambles, particularly in channel conditions where conventional methods are inefficient, leading to increased complexity and reduced performance in preamble detection and timing alignment.
Innovation Solution
The approach involves transmitting a message to a connected terminal device to perform a random access procedure, receiving the preamble signal as training input, and using this input to train a detection process for detecting preambles from unconnected devices, enabling improved preamble detection and timing advance correction through neural network-based methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional preamble detection methods are used, then the detection process is simple, but detection accuracy deteriorates in challenging channel conditions
Solution Approach 1:
The patent applies preliminary action by collecting training data from connected terminal devices before actual preamble detection. The access node gathers preamble signals from known devices, processes them through the neural network in advance, and uses this pre-trained model for detecting preambles from unconnected devices. This preliminary training phase improves detection accuracy for challenging channel conditions without increasing real-time detection complexity.
2Measurement precision
If neural network-based detection is applied to all terminal devices, then detection accuracy improves, but system complexity and resource consumption increase
Solution Approach 1:
The patent applies local quality by differentiating the detection approach based on device connection status. For connected terminal devices, conventional detection methods are used. For unconnected terminal devices operating in challenging channel conditions, the neural network-based detection is applied. This selective application optimizes detection accuracy where needed while avoiding unnecessary complexity for devices in good conditions.
Solution Approach 2:
The patent applies self-service by having connected terminal devices contribute their preamble signals as training data for the neural network. These connected devices essentially help train the system to detect preambles from unconnected devices, reducing the need for extensive external training data and enabling the system to improve its own detection capabilities.
3Measurement precision
If more training data is collected from connected devices, then detection performance improves, but signaling overhead and procedure time increase
Solution Approach 1:
The patent applies continuity of useful action by collecting training data during normal random access procedures. Instead of dedicating separate time resources solely for training data collection, the system utilizes preamble signals from connected devices as they naturally occur during ongoing random access operations. This continuous collection approach accumulates training data without interrupting normal system operation or requiring additional dedicated time resources.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
This document discloses a solution for preamble detection. According to an aspect, a method comprises: transmitting a message to a first terminal device connected to an access node, the message instructing the terminal device to carry out a random access procedure towards the access node; receiving a signal from the first terminal device during the random access procedure, the signal comprising a random access preamble; using the received signal as a training input in a random access preamble detection process; and using the random access preamble detection process trained with the training input in detection of a random access preamble from a second terminal device not connected to the access node.