Mesh Wi-Fi Client Steering Using ML and Wi-Fi RTT
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wireless networks experience reduced throughput and reliability when client devices are far from routers, especially with multiple devices in use, leading to bandwidth issues and suboptimal connections.
Innovation Solution
Implementing a machine learning model in a wireless mesh network that uses RSSI, distance, and time/day metrics to steer client devices to the most optimal access point for improved throughput, leveraging Wi-Fi RTT for location accuracy and machine learning to predict device movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If mesh Wi-Fi is deployed to extend coverage, then wireless coverage area is improved, but network throughput deteriorates when client devices are far from the router
Solution Approach 1:
The patent divides the wireless network into multiple mesh nodes (routers and access points) distributed throughout the coverage area. Each node independently manages client associations and performs local intelligence operations, segmenting the network control function to maintain throughput efficiency while extending coverage.
Solution Approach 2:
The system performs preliminary client steering by predicting future client locations and proactively associating clients with optimal access points before they move. This preliminary action prevents throughput degradation by ensuring clients are already connected to the best available node before signal quality deteriorates.
2Adaptability or versatility
If multiple client devices are connected to the network, then network coverage utility is improved, but bandwidth is spread thin and access speed slows down
Solution Approach 1:
The patent implements local quality by allowing different mesh nodes to serve different client devices based on their specific conditions. Each node optimizes its local client associations independently, ensuring that high-bandwidth clients are served by nodes with available capacity while maintaining overall network adaptability.
Solution Approach 2:
The system dynamically adjusts client associations across mesh nodes based on real-time network conditions, client locations, and bandwidth availability. This dynamic reassociation ensures that access speed is maintained by continuously optimizing the distribution of clients across available network resources.
3Device complexity
If traditional client association methods are used, then device complexity is reduced, but reliability of connection deteriorates when clients are far from router
Solution Approach 1:
The patent implements self-service by enabling mesh nodes to autonomously perform client steering decisions using local intelligence and captured signals. Nodes automatically associate clients with optimal access points without requiring centralized server intervention, maintaining connection reliability while minimizing system complexity.
Solution Approach 2:
The system uses feedback from captured signals (RSSI, RTT) and client behavior patterns to continuously optimize client associations. Mesh nodes monitor network conditions and automatically adjust associations based on this feedback, improving connection reliability without adding significant system complexity.
4Productivity
If machine learning model is implemented for client steering, then network performance is improved, but use of energy increases
Solution Approach 1:
The patent applies partial action by implementing machine learning capabilities selectively at edge mesh nodes rather than uniformly across the entire network. Only nodes that benefit from intelligent steering perform ML operations, reducing overall energy consumption while maintaining network performance improvements where needed.
Solution Approach 2:
The system extracts and uses only the essential features from captured signals (RSSI, RTT, basic client behavior) for machine learning operations, rather than processing complete data sets. This extraction approach maintains network performance benefits while minimizing the computational energy required for ML inference.
Data Source
AI summary
Apparatuses, methods, and systems for client steering in mesh networks are disclosed. A first access point (AP) of a network sends a quality of service (QoS) data packet to a client device. A received signal strength indication (RSSI) of the client device is captured using the QoS data packet. A distance between the AP and client device is determined using a Wi-Fi round trip time. A time and day of week is determined. Using machine learning, a second AP is identified for steering the client device to, based on the RSSI, the distance, and the time and day of week. A machine learning model is trained to steer each client device to a respective AP for increasing a throughput of the network based on features extracted from client behavior of the client devices. The second AP is connected to the client device.