Automated Access Point Placement via Graph-Based Reinforcement Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AP coordination methods require prior knowledge of the installation area, including floor plans and obstacle details, to optimize wireless performance, which can be complex and costly, and do not automate the placement of Access Points (APs) for optimal network performance without such information.
Innovation Solution
The use of a graph network and reinforcement learning to translate a physical space into a logical space, predicting signal strengths and evaluating AP placements without requiring specific knowledge of the location, structure, or obstacle details, allowing for automated and optimized AP placement for maximum wireless performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional AP coordination methods use floor plans and obstacle details to optimize wireless performance, then network performance is improved, but system complexity and implementation cost increase
Solution Approach 1:
The patent creates a virtual copy of the physical environment called 'logical space' that replicates the essential characteristics of the installation area without requiring actual floor plans or obstacle information. This virtual model allows the reinforcement learning algorithm to train and evaluate AP placements as if it were the real physical space, eliminating the need for complex site surveys and manual input while maintaining optimization accuracy
Solution Approach 2:
The patent introduces a graph network as an intermediary layer between the physical AP placements and the reinforcement learning algorithm. This graph network automatically translates physical coordinates and environmental features into a structured logical representation, serving as a mediator that bridges the gap between physical reality and the AI model without requiring manual intervention or complex data processing
2Manufacturing precision
If manual site surveys and floor plan analysis are performed to determine optimal AP placement, then placement accuracy is improved, but time consumption and cost increase
Solution Approach 1:
The reinforcement learning algorithm performs self-learning and self-optimization by training on the virtual logical space environment. The system automatically evaluates different AP placement scenarios and learns optimal strategies without requiring human experts to conduct site surveys or analyze floor plans. The algorithm improves its placement accuracy through iterative learning, making the system self-sufficient and eliminating time-consuming manual processes
Solution Approach 2:
The patent performs preliminary training of the reinforcement learning model in the virtual logical space before actual deployment. The algorithm learns from simulated environments and pre-evaluates placement strategies, so that when deployed in the real world, it can immediately provide accurate AP placement recommendations without requiring time-consuming on-site surveys or manual analysis of the actual installation area
3Measurement precision
If prior knowledge of installation area is required for AP coordination, then optimization accuracy is improved, but ease of deployment deteriorates
Solution Approach 1:
The logical space virtual environment serves multiple functions: it acts as both the training environment for the reinforcement learning algorithm and the evaluation space for testing placement strategies. This universal virtual model can represent any physical installation area regardless of its specific characteristics, making the system universally applicable to different environments without requiring environment-specific customization or prior knowledge
Solution Approach 2:
The patent transforms the problem from requiring detailed physical parameters (floor plans, obstacle locations, wall materials) to using abstract logical space parameters. By changing the representation from concrete physical measurements to virtual logical coordinates and graph structures, the system maintains optimization accuracy while eliminating the need for complex input data collection and processing
Data Source
AI summary
AP coordination, and more specifically intelligent AP coordination using a graph network and reinforcement learning may be provided. AP coordination may include translating a physical space into a logical space, wherein the physical space is being evaluated for AP coordination. A machine learning process may predict signal strengths of signals sent by one or more Access Points (APs) and received by one or more Stations (STAs), wherein the machine learning process uses the logical space, and wherein each STA is in a location of the physical space. One or more AP placements may be evaluated based on the signal strengths, and a recommended AP placement may be determined based on the evaluation.