Residential Router Traffic Inference for User Presence Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems are not adapted to determine user presence in residential spaces, relying on GPS data and network connectivity, which are inadequate for residential environments.
Innovation Solution
A machine learning system that collects network usage data from residential network routers and processes it using a machine learning algorithm to infer user presence, enabling accurate classification of user devices as stationary or non-stationary and detecting transitions within the residential space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS data and network connectivity methods are used to determine user location, then user presence can be identified in public spaces, but these methods are inadequate for residential environments
Solution Approach 1:
The patent applies local quality by implementing different detection methods for different environments. For residential spaces, it uses network router data collection and machine learning algorithms tailored to residential network behavior patterns, whereas public spaces can use GPS and network connectivity methods. This environment-specific adaptation resolves the contradiction between reliability and adaptability.
Solution Approach 2:
The system changes parameters by transitioning from GPS coordinates and network connectivity status to network usage data parameters such as data transfer volumes, packet counts, and protocol distributions. The machine learning model processes these transformed parameters to infer user presence, making the system adaptable to residential environments where GPS is unavailable.
2Measurement precision
If network usage data is collected and processed through machine learning algorithms, then accurate user presence inference in residential spaces is achieved, but system complexity increases
Solution Approach 1:
The patent introduces a machine learning model as an intermediary between raw network usage data and user presence determination. The model includes multiple processing layers (input layer, hidden layers, output layer) that transform complex router data into interpretable presence predictions, managing system complexity while maintaining high inference accuracy.
Solution Approach 2:
The system implements self-service by having the machine learning model automatically train and adapt to each household's unique network behavior patterns. The model learns from historical network data and continuously improves its predictions without requiring manual configuration or intervention, reducing operational complexity despite the sophisticated algorithms involved.
Data Source
AI summary
As described herein, a machine learning system, method, and computer program are provided for inferring user presence in a residential space. In use, network usage data is collected from a residential network router operating in a residential space. Additionally, the network usage data is processed by a machine learning algorithm to infer whether a user is present in the residential space. Further, the inference is output for performing one or more related actions.