Memory Cell Subdivision for Variable Condition Data Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for storing learning values in vehicle control units face challenges in differentiating between varying conditions due to limited memory capacity, leading to reduced accuracy and increased storage time as a result of overwriting data in shared cells and requiring a large number of subdivisions.
Innovation Solution
The method involves determining a cell corresponding to each variable condition and subdividing it into smaller cells, allowing for precise storage of learning values, reducing the number of learning times, and establishing prohibitions on further cell division to optimize storage efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a memory having a very large capacity is used to store learning values under different conditions, then the accuracy of storing learning values is improved, but the number of learning times increases geometrically, which significantly reduces the speed of storing the learning values
Solution Approach 1:
The patent divides the memory area into multiple virtual cells (e.g., 4×4 or 6×6 grid) to segment the storage space. This allows learning values under different conditions to be stored in different cells, improving accuracy without requiring a very large memory capacity. The segmentation enables efficient use of limited memory while maintaining the ability to distinguish between different operating conditions.
Solution Approach 2:
The patent introduces a virtual dimension by dividing the memory into a grid of cells, transforming a one-dimensional storage problem into a two-dimensional organization. This dimensional change allows for more efficient indexing and retrieval of learning values based on multiple conditions (e.g., vehicle speed, acceleration) without proportionally increasing memory size or learning time.
2Quantity of substance
If the same learning cell in the memory is shared under slightly different conditions, then the memory capacity is utilized efficiently, but different learning values calculated under different conditions may be stored in the same learning cell, which significantly reduces the accuracy
Solution Approach 1:
The patent applies local quality by assigning specific conditions to specific cells within the memory grid. Each cell is optimized to store learning values for particular operating conditions (e.g., specific vehicle speed ranges, acceleration levels). This localized assignment ensures that learning values under different conditions are stored in appropriate cells, maintaining accuracy while efficiently utilizing memory capacity.
Solution Approach 2:
By segmenting the memory into multiple virtual cells, the patent creates distinct storage locations for different operating conditions. This segmentation prevents different learning values from being stored in the same cell, thereby maintaining accuracy while still achieving efficient memory utilization through the structured grid organization.
Data Source
AI summary
A method for storing data in a memory may include when the data may be obtained under a variable condition, determining a cell corresponding to an area of the variable condition from an entire memory area, storing the data in the cell and dividing the cell storing the data into a plurality of cells, and whenever a new data may be obtained under a new variable condition, determining a new cell corresponding to an area of the new variable condition under which the new data may be obtained, from the plurality of cells of the entire memory area and repeating storing of the new data and dividing of the new cell.