Ultrasonic Access Control Reader with ML Intent Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In access control systems with co-located sound-based access control devices, authentication messages can inadvertently unlock multiple secure areas or assets, compromising security by failing to accurately determine the intended access control device of interest.
Innovation Solution
A method using a first access control device to collect environmental information, including ultrasonic, infrared, and exit sensor data, and employing a machine learning model, such as a convolutional neural network, to determine the access intention and selectively grant access to the intended secure area or asset, thereby preventing incorrect unlocking of co-located devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If co-located access control devices are used to control multiple secure areas, then the coverage and convenience of access control is improved, but the security is worsened because authentication messages may inadvertently unlock multiple secure areas
Solution Approach 1:
The patent segments the access control system into multiple independent reader devices, each associated with a specific secure area. Each reader device independently processes authentication messages and determines whether to grant access based on its own environmental context, preventing one reader from inadvertently unlocking another secure area.
Solution Approach 2:
The patent applies local quality by having each reader device make access control decisions based on its local environmental information (such as proximity to the secure area, ambient acoustic characteristics). This ensures that access is granted only to the intended secure area based on the specific local context rather than a centralized uniform decision.
2Ease of operation
If acoustic broadcasting is used for authentication, then the ease of operation is improved, but the precision of access control is worsened because it cannot accurately determine the intended secure area
Solution Approach 1:
The patent implements feedback mechanisms where reader devices continuously monitor and report environmental information (acoustic field characteristics, spatial proximity data) back to the authentication system. This feedback loop enables the system to accurately determine which secure area the user intends to access based on real-time environmental conditions, improving precision while maintaining ease of operation.
Solution Approach 2:
The patent introduces environmental information as an intermediary factor that mediates between the acoustic authentication signal and the access control decision. This intermediary (environmental context) provides the additional precision needed to disambiguate which secure area is the intended target, while the acoustic broadcasting remains the primary convenient authentication method.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances security by accurately identifying the intended access control device and restricting access to the correct secure area or asset, preventing unauthorized access to nearby secure areas or assets.
Implementation Method 1
a first reader device configured to control access to a first secure area and/or asset via first ultrasound communications
Implementation Method 2
the first input information including power information, foot information, infrared (IR) information
Data Source
AI summary
Example implementations include a method, system, and computer-readable medium, comprising collecting environment information by a first reader device configured to control access to a first secure area via ultrasound communications. The implementations further include determining first input information based on the environment information, the first input information. Additionally, the implementations further include determining, via a machine learning model, access intention information identifying the first secure area or a second secure area as an object of interest based on the first input information and second input information, wherein the second input information is associated with a second reader device that controls access to the second secure area and is co-located with the first reader device. Additionally, the implementations further include providing, based on the access intention information, access to one of the first secure area or the second secure area.