Autonomous Vehicle Memory Mapping Using Section Access Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for assigning sections of tasks to storage devices in MCU programs, such as those used in vehicle control systems, often fail to account for the unique characteristics of various programs, leading to inefficiencies in task execution time, power consumption, and heat issues, and require manual intervention by developers, which is time-consuming and costly.
Innovation Solution
A vehicle control apparatus that generates profiling data based on access counts of sections, determines priorities, and sequentially assigns these sections to storage devices based on priority, considering storage capacity and access cycles, minimizing manual intervention and optimizing memory mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a compiler automatically assigns sections to storage devices based on general program characteristics, then the memory mapping process is efficient and automated, but it cannot reflect the specific characteristics of various types of programs leading to suboptimal performance
Solution Approach 1:
The system performs preliminary profiling of program execution to collect access count data for each section before final memory mapping. This preliminary analysis enables the automated system to adapt to specific program characteristics by identifying which sections are accessed most frequently, allowing optimization based on actual program behavior rather than general assumptions.
2Adaptability or versatility
If a developer manually analyzes program characteristics and creates memory maps, then the memory mapping reflects specific program characteristics and optimizes performance, but it requires significant time and developer intervention
Solution Approach 1:
The system enables self-service memory mapping by automatically profiling program execution and generating optimized memory maps without requiring developer intervention. The control device executes the program, collects access count data, determines priorities automatically, and generates the memory map configuration, eliminating the time-consuming manual analysis process while maintaining adaptability to specific program characteristics.
3Productivity
If sections are assigned to storage devices without considering access patterns, then the memory mapping process is simple and fast, but task execution time and power consumption are not optimized
Solution Approach 1:
The system uses feedback from actual program execution to optimize memory mapping. By profiling the program and collecting access count data for each section, the system gains feedback about which sections are accessed most frequently. This feedback is then used to assign high-priority sections to faster storage devices, optimizing task execution time and power consumption while maintaining efficient memory mapping through automated priority-based assignment.
4Duration of action of moving object
If high-priority sections are assigned to storage devices with faster access cycles, then task execution time and power consumption are reduced, but storage capacity constraints must be carefully managed
Solution Approach 1:
The system changes the parameter of section assignment from static compiler-based or manual methods to dynamic priority-based assignment. By calculating access counts and determining priorities as variables, the system can flexibly assign sections to storage devices based on actual access patterns. This parameter change enables optimization of execution time while managing storage capacity constraints through automated priority comparison and selection algorithms.
Data Source
AI summary
An apparatus for controlling autonomous driving of a vehicle is introduced. The apparatus may generate profiling data comprising an access count of each of a plurality of sections, wherein the plurality of sections are associated with at least one task performed for an operation control of the vehicle. The apparatus may determine, based on the profiling data, a priority for each of the plurality of sections. Further, the apparatus may assign, based on the priority, the plurality of sections to one or more of a plurality of storage devices included in the memory, output, based on the sequentially assigned plurality of sections, a signal for the operation control of the vehicle, and control, based on the signal, an operation of autonomous driving of the vehicle.