Vehicle Access Verification Using Gait Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle access verification systems rely on physical objects or sensitive biometric data, which can be misused or compromised, necessitating a method that identifies authorized individuals without these requirements.
Innovation Solution
An access verification system utilizing non-visual sensors to detect and analyze motion characteristics, such as gait, to authenticate individuals, employing machine learning for accurate identification and allowing access without physical objects or sensitive biometric data, using existing vehicle sensors and machine learning to create unique user profiles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If physical objects or sensitive biometric data are used for access verification, then identification accuracy is improved, but security risk increases due to potential misuse or compromise
Solution Approach 1:
The patent replaces traditional mechanical access verification systems (physical keys, cards) and sensitive biometric systems (fingerprints, facial recognition) with a motion-based verification system using non-visual sensors. The system detects and analyzes motion characteristics such as gait patterns, walking speed, and movement rhythm to identify authorized individuals without requiring physical objects or sensitive biometric data, thereby maintaining identification accuracy while reducing security risks associated with data compromise
Solution Approach 2:
The patent changes the verification parameter from static physical objects or sensitive biometric data to dynamic motion characteristics. By monitoring parameters such as walking speed, step frequency, gait pattern, and movement rhythm over time, the system creates a behavioral biometric profile that is difficult to replicate or compromise, thus improving security while maintaining identification precision
2Object-affected harmful factors
If non-visual sensors and motion analysis are used, then security risk is reduced, but measurement precision may worsen due to difficulty in detecting and measuring motion characteristics
Solution Approach 1:
The patent combines multiple non-visual sensors (such as ultrasonic sensors, infrared sensors, or radar) to detect and measure motion characteristics simultaneously. By merging data from multiple sensor sources and analyzing multiple motion parameters (walking speed, gait pattern, step frequency, movement rhythm) together, the system achieves high measurement precision in identifying motion characteristics while maintaining the security benefits of non-visual detection
Solution Approach 2:
The system continuously monitors motion characteristics and compares detected patterns against stored profiles of authorized individuals. The feedback mechanism refines motion detection accuracy by learning from multiple detection attempts and adjusting recognition thresholds, thereby improving measurement precision over time while maintaining security through consistent verification protocols
3Measurement precision
If machine learning is employed for gait recognition, then identification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent implements a preliminary registration phase where motion characteristics of authorized individuals are captured and stored as reference profiles before actual access verification. During registration, the system collects and analyzes gait patterns, walking speed, and movement rhythms to create baseline profiles. This preliminary action simplifies subsequent verification operations, as the system only needs to compare real-time motion data against pre-established profiles rather than performing complex analysis during each access attempt
Solution Approach 2:
The system creates simplified digital representations (copies) of complex motion patterns by extracting key特征 parameters such as walking speed, step frequency, and gait rhythm. These copied motion signatures are stored as compact data structures that can be quickly compared during verification, reducing the computational complexity of the recognition system while maintaining high identification accuracy
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 system achieves high accuracy (up to 98%) in identifying registered individuals based on their gait patterns, providing secure access to vehicles without the need for physical keys or sensitive biometric data, while maintaining low energy consumption and being retrofittable to existing vehicles.
Implementation Method 1
The at least one sensor is configured to non-visually detect a motion within the predefined periphery of the vehicle
Data Source
Figure 1~3
AI summary
The disclosure relates to an access verification system (100) for a vehicle (200), comprising a monitoring unit (102) configured to detect a motion near the vehicle (200), comprising at least one sensor (1021) configured to detect a motion near the vehicle (200); and a sensor data processing unit (1022) configured to determine if the detected motion corresponds to at least one motion-characteristic of a person (112) approaching the vehicle (200), a verification unit (104) configured to be activated, when the motion detected by the monitoring unit (102) is a motion-characteristic of a person (112), and when the detected motion is equal to or larger than a predetermined motion- threshold, and configured to determine if the person (112) approaching the vehicle (200) corresponds to a registered person by comparing the at least one characteristic of the person (112) detected by the monitoring unit (102) with at least one pre-registered motion-characteristic of the at least one registered person, and a processing unit (108) configured to provide the person (112) access to the vehicle (200). Further, the disclosure relates to an access verification method (300).