Vehicle Control Learning Data Rewrite for Post-Update Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
When a vehicle control program is updated, using pre-update learning values can lead to inappropriate vehicle control, particularly in the initial stages after the update, due to potential disruptions in hydraulic-pressure command values and shifting shocks.
Innovation Solution
A vehicle control apparatus that includes a learning-data rewrite portion to adjust correction values from pre-update to post-update values, ensuring the post-update values have the same sign as pre-update values but with a smaller absolute value, and adjusts based on the degree of progress of the learning operation, thereby reducing the influence of pre-update values on post-update control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pre-update learning values are maintained after control program update, then data continuity is preserved, but vehicle control appropriateness deteriorates in initial stages post-update
Solution Approach 1:
The patent applies parameter changes by modifying the correction values stored in memory. Specifically, when a control program update is detected, the system changes the correction values from pre-update values to post-update values through a learning operation. This parameter transformation ensures that the correction values are appropriate for the updated control program while maintaining the learning mechanism's continuity, thereby resolving the contradiction between data continuity and control appropriateness.
Solution Approach 2:
The system performs preliminary action by detecting control program updates and initiating learning operations before normal control resumes. The learning-data rewrite portion proactively rewrites correction values when an update is detected, and the learning operation executes in advance to generate appropriate post-update correction values. This preliminary processing ensures that control appropriateness is maintained from the start of post-update operation.
2Measurement precision
If pre-update correction values are used after control program update, then learning data continuity is maintained, but hydraulic-pressure command accuracy deteriorates
Solution Approach 1:
The patent introduces an intermediary mechanism - the learning operation - that transforms pre-update correction values into post-update correction values. The learning-data rewrite portion acts as an intermediary component that detects control program updates and triggers the learning operation. This intermediary process ensures that correction values are appropriately adapted to the updated control program, maintaining hydraulic-pressure command accuracy while managing the complexity through a structured transformation process.
Solution Approach 2:
The system employs feedback through the learning operation that continuously monitors and adjusts correction values. When a control program update is detected, the learning operation executes to generate appropriate correction values by comparing actual system behavior with target values. This feedback mechanism ensures that correction values remain accurate for the updated control program while providing a systematic approach to managing the complexity of correction value adaptation.
3Reliability
If learning operation is restarted after control program update, then control appropriateness is ensured, but time for learning is extended
Solution Approach 1:
The patent applies partial action by selectively restarting only the necessary learning operations after control program updates. The learning-data rewrite portion identifies which correction values need to be updated and triggers learning operations only for those specific parameters. This partial approach ensures control appropriateness for updated functions while avoiding unnecessary extension of learning duration for unaffected parameters, thereby resolving the contradiction between reliability and learning duration.
Data Source
AI summary
A vehicle control apparatus for controlling a vehicle through a control program using at least one parameter. The at least one parameter is corrected by respective at least one correction value that is obtained after start of execution of a learning operation. The vehicle control apparatus includes: a learning-data storage portion configured to store, as learning data, the at least one correction value obtained after the start of the execution of the learning operation; and a learning-data rewrite portion configured, when the control program is updated, to execute a rewriting operation for rewriting the at least one correction value as the learning data from a pre-update correction value to a post-update correction value, such that the post-update correction value has the same sign as the pre-update correction value, and an absolute value of the post-update correction value is smaller than an absolute value of the pre-update correction value.