Edge Server-Assisted Sensor Data Processing for 5G Uplink Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication networks, particularly 5G cellular networks, face challenges with asymmetric data rates, where the uplink data rate is significantly lower than the downlink, hindering the transfer of large amounts of data from communication end nodes to edge servers, especially in real-time applications like autonomous driving, which can negatively impact network services.
Innovation Solution
A communication device equipped with a processor that executes real-time applications, determines confidence values of processed data, and sends data to an edge server only if the confidence value is below a threshold, allowing the edge server to process and return results, thereby optimizing data transfer and leveraging edge computing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If large amounts of sensor data are transferred from communication end nodes to edge server, then real-time processing capability is improved, but uplink data rate becomes insufficient and network services are negatively affected
Solution Approach 1:
The patent applies local quality by enabling communication end nodes to perform local processing of sensor data using machine learning models. Instead of transferring all raw sensor data to the edge server, only selectively processed data or results are transmitted. This resolves the contradiction by improving real-time processing capability locally while reducing the burden on the limited uplink data rate.
Solution Approach 2:
The patent segments the data processing task into two parts: initial processing and filtering is performed at the communication end node, while more complex analysis is performed at the edge server. This segmentation allows critical real-time processing to occur locally without overwhelming the uplink channel, while still leveraging edge server resources for comprehensive analysis.
2Measurement precision
If all sensor data is sent to edge server for processing, then processing accuracy is improved, but network bandwidth is overwhelmed and other services are affected
Solution Approach 1:
The patent applies preliminary action by having communication end nodes perform preliminary processing of sensor data before transmission to the edge server. Machine learning models on end nodes pre-process data, extract relevant features, or filter out unnecessary information, thereby improving processing accuracy at the edge server while significantly reducing the quantity of data transmitted over the network.
Solution Approach 2:
The patent extracts only the essential or most valuable data from the complete sensor data set for transmission to the edge server. By using machine learning models to identify and extract critical information, the system maintains high processing accuracy while minimizing network bandwidth consumption.
3Measurement precision
If confidence value threshold is set low, then more data is processed by edge server improving accuracy, but more network traffic is generated
Solution Approach 1:
The patent applies dynamics by making the confidence value threshold adjustable and adaptive rather than fixed. The threshold can be dynamically modified based on network conditions, data criticality, and service requirements. This allows the system to optimize the balance between processing accuracy and network traffic volume in real-time, sending more data to the edge server when accuracy is critical and fewer data when network bandwidth is constrained.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a communication device (120) for communication in a communication network (100) having an edge server (140). The communication device (120) comprises: a sensor (121) configured to provide sensor data; a processor (123) configured to execute a real-time application, wherein the real-time application is configured to process the sensor data and to provide a result on the basis of processing the sensor data, wherein the processor (123) is further configured to determine a confidence value of the result; and a communication interface (125) configured to send, in case the confidence value of the result provided by the real-time application is smaller than a threshold value, the sensor data to the edge server (140) and to receive from the edge server (140) a further result based on processing the sensor data by the edge server (140).