RF Plan Generation Using Predictive Models for Wireless Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual RF planning for wireless network deployments is error-prone and requires frequent site visits, leading to suboptimal network performance due to changes in deployment areas and user density.
Innovation Solution
A dynamic network management system employing a trained predictive model to determine the optimal number and layout of access points based on deployment area characteristics, using IoT devices and machine-learning algorithms to generate and adjust RF plans efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual RF planning is used to survey deployment areas, then RF plans can be generated, but errors occur and frequent site visits are required
Solution Approach 1:
The patent replaces manual mechanical surveying processes with automated image processing and machine learning algorithms. The system captures images of deployment areas, automatically extracts features (walls, doors, windows, furniture), and generates RF plans without requiring manual site visits, thereby improving reliability and reducing time loss.
Solution Approach 2:
The patent creates digital copies of the physical deployment area by processing images to generate virtual models. These digital models include extracted features and simulated RF coverage that accurately represent the physical space, enabling reliable RF planning without repeated physical site visits.
2Productivity
If manual surveying is performed to identify AP layout, then RF plans can be created, but the process is error-prone and time-consuming
Solution Approach 1:
The patent replaces manual surveying and analysis with automated computer vision and machine learning systems. The system automatically processes images, extracts environmental features, simulates RF coverage, and generates optimized AP layouts, dramatically improving productivity while maintaining or enhancing network performance reliability.
Solution Approach 2:
The system performs self-service by automatically capturing images, processing them through machine learning models, extracting deployment area features, simulating RF coverage patterns, and generating optimized RF plans without human intervention, thereby increasing productivity while ensuring reliable network performance through algorithmic optimization.
3Adaptability or versatility
If RF plans are generated without considering deployment changes, then initial plans can be created quickly, but they become suboptimal when user density or layout changes
Solution Approach 1:
The patent implements dynamic RF planning by enabling the system to process new images and regenerate RF plans when deployment areas change. The machine learning models can adapt to new furniture arrangements, wall modifications, or user density changes by reprocessing images and updating AP layouts accordingly, making the system adaptable without requiring extensive manual re-surveying.
Solution Approach 2:
The system performs preliminary actions by capturing and processing images of the deployment area in advance, extracting features and generating baseline RF plans before actual deployment. This allows for quick updates when changes occur, as the system already has processed data and models ready to adapt to new conditions, reducing the time needed for plan updates.
Data Source
AI summary
The present disclosure provides an improved computing system to determine a suitability of a RF plan and a dynamic wireless network system. The method includes receiving uploads of RF plans associated with wireless network deployments which provided wireless network capability throughout each respective deployment area. Further, storing the uploaded RF plans with in a network storage cloud. In addition, training a predictive model with the stored uploaded RF plans. The dynamic wireless network system disclosed herein includes a plurality of internet-of-things devices within a deployment area. The system further includes at least one network device communicatively coupled to each of the plurality of internet-of-things devices, wherein the at least one network device has access to a trained predictive model that is to determine a number of access points and a layout of the access points within the deployment area based on a set of characteristics associated with the deployment area.