Autonomous Driving Route Selection for Log Memory Capacity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The limited memory capacity in vehicles equipped with autonomous driving systems using machine learning models leads to a situation where log data fills the available space, preventing new log data from being stored during autonomous driving control, as existing methods do not effectively manage data storage and transmission.
Innovation Solution
The autonomous driving system predicts the fluctuation amount of log data for multiple route candidates and selects the route that maximizes the future free space in the memory device, ensuring sufficient storage by choosing the route with the highest predicted free space value at the destination, thereby preventing memory overflow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If log data is continuously stored in the memory device during autonomous driving control, then the data quantity for verification increases, but the free space of the memory device decreases and may be filled up
Solution Approach 1:
The system performs preliminary prediction of log data generation amounts for each candidate route before execution. By calculating expected data quantities in advance and selecting routes that maximize future free space, the system prevents memory overflow before it occurs, ensuring continuous logging capability throughout the autonomous driving task.
2Quantity of substance
If the memory device capacity is limited, then storage cost is reduced, but log data may not be stored when free space is filled
Solution Approach 1:
The system performs preliminary prediction of log data generation amounts for each candidate route before execution. By calculating expected data quantities in advance and selecting routes that maximize future free space, the system prevents memory overflow before it occurs, ensuring continuous logging capability throughout the autonomous driving task.
3Extent of automation
If autonomous driving control is executed using machine learning models, then driving automation is improved, but verification of the models requires substantial log data storage
Solution Approach 1:
The system performs preliminary prediction of log data generation amounts for each candidate route before execution. By calculating expected data quantities in advance and selecting routes that maximize future free space, the system prevents memory overflow before it occurs, ensuring continuous logging capability throughout the autonomous driving task.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An autonomous driving route along which autonomous driving control of a vehicle (1) is executed is set. In the setting of the autonomous driving route, at least two candidates of a route (RTn) from a current location to a destination of the vehicle are set. Next, for each of the at least two candidates, a fluctuation amount of log data that fluctuates in a memory device (120) during a travel of the vehicle along the candidate of the route. Next, a future value of a free space in the memory device at which the vehicle arrives at a destination of the candidate of the route (QC_RTn (Tds)) for each of the at least two candidates based on the fluctuation amount of the log data. Then, one candidate of the route from the at least two candidates of which the future value of the free space is the maximum is selected as the autonomous driving route.