Floor Plan Material Mapping with Machine Learning for Radio Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Indoor cellular network design faces challenges due to the lack of accurate material information for obstacles in floor plans, leading to inefficiencies and increased time in onsite investigations by field technicians, which affects the precision of radio signal predictions.
Innovation Solution
A machine learning-based approach using a U-Net model in a conditional generative adversarial network (cGAN) to estimate obstacle materials from floor plans, generating a material map and obstacle map to automate the augmentation of floor plans with precise material information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If field technicians perform onsite investigation to identify obstacle materials, then material information accuracy is improved, but time consumption and process complexity increase
Solution Approach 1:
The patent uses a U-Net machine learning model to generate a material map that copies and predicts material information from floor plan images. Instead of physical onsite investigation, the system creates a digital copy of material distribution by training the model on paired datasets of floor plans and their corresponding material maps, then uses this trained model to automatically predict materials for new floor plans.
Solution Approach 2:
The patent replaces the mechanical process of onsite investigation with an automated image processing system. The U-Net model processes floor plan images through convolutional neural network layers, substituting human field technicians with an automated computational system that extracts material information directly from visual data.
2Measurement precision
If field technicians perform onsite investigation to identify obstacle materials, then material information accuracy is improved, but process complexity increases
Solution Approach 1:
The system creates a material map that is a digital copy of the actual material distribution in the building. By training the U-Net model on paired datasets, the system learns to copy material information patterns from floor plans, eliminating the need for complex manual investigation processes.
Solution Approach 2:
The floor plan image itself serves as the source of material information. The U-Net model enables the system to extract material data directly from the visual content of the floor plan without requiring external onsite investigation, making the process self-sufficient and reducing overall complexity.
3Loss of information
If obstacle segmentation is performed by eyeballing in floor plans, then material information can be obtained, but accuracy and efficiency deteriorate
Solution Approach 1:
The patent replaces manual eyeballing and segmentation with an automated U-Net machine learning model. The model processes the entire floor plan image through multiple convolutional layers and generates a complete material map automatically, substituting human visual inspection with computational image analysis that is both faster and more consistent.
Solution Approach 2:
The U-Net model processes the floor plan in a continuous automated workflow from input image to output material map. The encoder-decoder architecture with skip connections maintains continuous feature flow throughout the network, enabling seamless processing without the intermittent human intervention required in manual segmentation.
Data Source
AI summary
Estimating materials of obstacles from a floor plan (310) is described. A floor plan is loaded (210) and preprocessed (215, 410). A material map (510) for the preprocessed floor plan is generated using a machine learning model for estimating materials of obstacles of the floor plan (220, 225). An obstacle map (610) is generated from the preprocessed floor plan that includes segmented obstacles as line segments. An augmented floor plan (710) is generated (225) based on a combination of the generated material map and the generated obstacle map, where the augmented floor plan identifies the estimated material of the segmented obstacles.