Autonomous Vehicle Behavior Visualization for Crossing Object Intent
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current user interfaces for autonomous vehicles (AVs) do not provide sufficient insight into the AV's planned behavior, leading to unnecessary manual overrides by human drivers or passengers who are unsure if the AV has considered objects or conditions in its surroundings.
Innovation Solution
A user interface system that visualizes the AV's planned behavior by classifying objects as asserting or yielding based on predicted pathways, using a user interface engine to generate images indicating whether the AV plans to yield or assert itself relative to objects, and displaying planned velocities and traffic light status.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If current user interfaces show only basic surrounding information and planned path, then the interface remains simple, but driver confidence and understanding of AV intentions are insufficient
Solution Approach 1:
The user interface is segmented into distinct functional zones: a first portion displaying basic surrounding information and a second portion displaying detailed planned behavior information. This segmentation allows the interface to provide comprehensive information about AV intentions (resolving the information loss) while maintaining organizational simplicity and avoiding overwhelming the user (managing interface complexity).
Solution Approach 2:
The interface transitions from a two-dimensional display of basic data to a multi-dimensional visualization that includes spatial relationships, temporal sequences, and hierarchical information layers. By presenting planned behavior information in an expanded dimensional format that shows relationships between AV actions and environmental objects, the system provides deeper insight without proportionally increasing interface complexity.
2Reliability
If the AV provides detailed information about its planned behavior, then driver confidence improves, but the processing and visualization requirements increase
Solution Approach 1:
The system extracts only the most relevant planned behavior information for display in the user interface, separating essential confidence-building data from comprehensive internal processing data. By extracting and displaying only key information about AV intentions and planned actions, the system builds driver confidence while minimizing computational resources dedicated to visualization.
Solution Approach 2:
The interface applies local quality by providing detailed planned behavior information specifically in the second portion where it is most needed for driver confidence, while keeping the first portion simple for quick reference. This localized enhancement of information quality builds reliability where it matters most without requiring high computational power across the entire interface.
3Measurement precision
If the AV classifies objects as asserting or yielding based on predicted pathways, then the visualization accuracy improves, but the classification complexity increases
Solution Approach 1:
Instead of analyzing complex interactions between multiple objects and predicting all possible outcomes, the system inverts the approach by classifying each environmental object individually as either asserting or yielding based on its relationship to the AV's planned path. This inversion simplifies the classification task while maintaining prediction accuracy by focusing on binary categorization rather than comprehensive multi-object analysis.
Data Source
AI summary
To visualize planned behavior of an autonomous vehicle (AV) traveling along a roadway, a user interface engine receives data describing a planned pathway of the AV along the roadway and object data describing an object having a predicted pathway crossing the planned pathway of the AV at a cross point. The user interface engine classifies the object either an asserting object or a yielding object based on a prediction of whether the object reaches the cross point before the AV or after the AV. The user interface engine generates an image that includes the planned pathway of the AV and the object in the environment of the AV. The image of the object indicates whether the object is classified as an asserting object or a yielding object.