Automated Driving Control With Explainable Travel State Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated driving systems using machine learning models face challenges in continuously making appropriate decisions across various travel scenarios, necessitating user understanding of the decision-making processes.
Innovation Solution
An automated driving system that includes an automated driving control unit using a machine learning model to decide on travel state changes based on a target route, vehicle position, map information, and sensor data, along with an information presentation unit that explains the reasons for these changes to the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a machine learning model is used to automatically decide travel state changes, then automation and decision-making speed are improved, but user understanding and trust in the system deteriorate
Solution Approach 1:
The patent introduces an information presentation unit as an intermediary between the automated driving control unit and the user. This mediator translates the internal machine learning decisions into externally comprehensible explanations, including the travel state change factor (reason) and the specific change content, thereby bridging the information gap without compromising automation
Solution Approach 2:
The system implements a feedback mechanism where user responses to presented information are received and used to adjust future information presentation. This creates a closed-loop system that continuously improves user understanding while maintaining automated decision-making, allowing the system to adapt to individual user needs and preferences
2Loss of information
If detailed information about decision factors is presented to the user, then user understanding and trust are improved, but information processing load and system complexity increase
Solution Approach 1:
The information presentation is segmented into distinct, manageable components: the travel state change factor (the reason) and the travel state change content (the action). This segmentation allows the complex decision-making process to be broken down into understandable parts, reducing cognitive load while maintaining completeness
Solution Approach 2:
The system dynamically adjusts information presentation parameters based on user feedback and usage history. By changing parameters such as information detail level, presentation timing, and format based on user responses, the system optimizes the balance between user understanding and system complexity without requiring fixed complex structures
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An automated driving system for deciding to change a travel state of a host vehicle using a machine learning model in an automated driving control process for the host vehicle includes an automated driving control unit configured to decide to change the travel state of the host vehicle using the machine learning model on the basis of a preset target route and position information of the host vehicle and map information, or a detection result of an external sensor of the host vehicle, and an information presentation unit configured to present a travel state change factor, which is a main cause of the change in the travel state, and the change in the travel state of the host vehicle to a user of the host vehicle when the automated driving control unit decides to change the travel state of the host vehicle.