Automatic Transmission Oil Pressure Learning for Clutch Wear Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing oil pressure adjustment methods for automatic transmissions fail to accurately account for changes in clutch or brake characteristics over time, leading to inadequate oil pressure adjustments.
Innovation Solution
An oil pressure learning method that uses reinforcement learning to adjust oil pressure based on detected input rotation speed, gear-shift time, heat generation, and torque changes, updating relational regulation data to optimize oil pressure command values for improved transmission performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If oil pressure is adjusted simply based on the blow amount of the input rotation speed, then the control method is simple, but the oil pressure cannot be appropriately adjusted according to the change in the characteristics of the automatic transmission
Solution Approach 1:
The patent implements feedback mechanisms by detecting multiple parameters (blow amount, gear-shift time, heat generation amount) and using them to update relational regulation data through reinforcement learning. The system continuously monitors transmission characteristics and adjusts oil pressure commands based on feedback from these detections, ensuring appropriate adaptation to clutch/brake wear while maintaining a structured control framework.
Solution Approach 2:
The patent changes multiple parameters simultaneously for comprehensive adjustment: it detects blow amount, gear-shift time, and heat generation amount; calculates torque variables based on input torque changes; and updates relational regulation data with learned values. This multi-parameter approach allows appropriate oil pressure adjustment according to transmission characteristics while avoiding oversimplification.
2Measurement precision
If reinforcement learning is used to update relational regulation data with multiple variables, then the oil pressure adjustment accuracy is improved, but the device complexity increases
Solution Approach 1:
The control device performs multiple functions using a unified reinforcement learning framework: it detects various parameters (blow amount, gear-shift time, heat generation), calculates torque variables, updates relational regulation data, and generates oil pressure commands. This multi-functional approach improves adjustment accuracy while avoiding the need for separate complex systems for each function.
Solution Approach 2:
The system uses itself to improve its own performance through reinforcement learning. The control device detects its own operational parameters, calculates performance metrics (blow amount, gear-shift time, heat generation), and autonomously updates its relational regulation data based on learned experiences. This self-service mechanism improves accuracy without requiring external complex adjustment systems.
3Adaptability or versatility
If the relational regulation data is updated frequently to adapt to changing transmission characteristics, then the adaptability is improved, but the stability of the control system may deteriorate
Solution Approach 1:
The patent updates relational regulation data based on learned values from reinforcement learning, which accumulates knowledge over multiple detection cycles. By preliminarily learning from past experiences and storing relational regulation data, the system adapts to transmission characteristics changes while maintaining stability through established learned patterns rather than making immediate abrupt changes.
Solution Approach 2:
The system dynamically balances adaptability and stability by using reinforcement learning to gradually update relational regulation data. The control device adapts to changing transmission characteristics (clutch/brake wear) over time through continuous learning, while the stability is maintained by building upon previously learned relational patterns rather than making sudden changes.
Data Source
AI summary
An oil pressure learning method of an automatic transmission includes acquiring a state of a vehicle in a state where relational regulation data is stored, supplying oil to the automatic transmission such that the value of the oil pressure is set to an oil pressure command value, calculating, as a specific variable, a variable representing an amount in which a detected input rotation speed exceeds a target input rotation speed, or the like, calculating a reward in a manner in which it has a larger value when the specific variable satisfies a criterion than when it does not satisfy the criterion, updating the relational regulation data by inputting, to an update mapping, the reward and the oil pressure command value, and calculating a torque variable having a value that is increased as an amount of change in an input torque is increased.