Remote Park-Assist Input Classification for Anomalous Gesture Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current remote park-assist systems face challenges in accurately verifying user inputs on mobile devices, leading to potential misinterpretation of intended and unintended gestures, which can result in incorrect vehicle control during autonomous parking.
Innovation Solution
A system that includes a mobile device with a touchscreen interface for remote parking, a vehicle autonomy unit, and an input classifier to verify inputs using machine learning algorithms, distinguishing between nominal and anomalous inputs, and sending notifications for non-compliant signals to ensure accurate vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If remote park-assist systems use touchscreen input without verification, then ease of operation is improved, but reliability deteriorates due to potential misinterpretation of gestures
Solution Approach 1:
The system performs preliminary classification of touchscreen inputs before executing remote parking commands. The input classifier analyzes touch patterns, duration, and position to verify whether the input is intentional before the vehicle autonomy unit acts on it, preventing misinterpretation of accidental gestures.
Solution Approach 2:
The system implements feedback by notifying the user when an input is received during remote parking operations. This allows the user to confirm or cancel the intended action, creating a verification loop that enhances reliability while maintaining ease of operation through intuitive interaction.
2Reliability
If the system verifies all inputs during remote parking, then reliability is improved, but device complexity increases due to input classification requirements
Solution Approach 1:
The input classifier is integrated into the existing vehicle autonomy unit, allowing it to self-verify inputs using the same processing resources already allocated for autonomous parking control. This eliminates the need for separate dedicated verification hardware, reducing overall system complexity while maintaining reliability.
3Measurement precision
If the system distinguishes between nominal and anomalous inputs, then measurement precision is improved, but difficulty of detecting and measuring increases
Solution Approach 1:
The input classifier uses multiple parameters including touch duration, touch position, swipe distance, and gesture pattern to distinguish between nominal and anomalous inputs. By analyzing combinations of these parameters rather than single metrics, the system achieves high classification accuracy while managing detection complexity through systematic parameter evaluation.
Data Source
AI summary
Method and apparatus are disclosed for interface verification for vehicle remote park-assist. An example vehicle system includes a mobile device and a vehicle autonomy unit. The mobile device includes a touchscreen and a controller. The controller is to present, via the touchscreen, an interface of a remote parking app and receive, via the touchscreen, an input responsive to the presentation of the interface. The vehicle autonomy unit receives an input signal from the mobile device and an input classifier coupled to the vehicle autonomy unit verifies the received input signal complies with an input classification.