Vehicle Control System Reinforcement Learning Throttle Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing vehicle control systems require extensive manual effort and man-hours to set the throttle valve opening degree of internal combustion engines appropriately based on accelerator pedal operation, leading to inefficiencies in adapting electronic equipment operations to vehicle states.
Innovation Solution
A vehicle control system comprising a memory, first and second processors, and a data analysis center that uses relationship definition data to update and optimize the operation of electronic equipment by acquiring sensor detection values, calculating rewards, and applying reinforcement learning to adjust the throttle valve opening degree and ignition timing based on vehicle states, reducing the need for manual expert intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual expert adjustment is used to set throttle valve opening degree, then the vehicle control can be precisely adapted to vehicle state, but extensive man-hours and manual effort are required
Solution Approach 1:
The control device automatically adjusts the throttle valve opening degree by itself based on detected vehicle states and reinforcement learning, eliminating the need for manual expert adjustment. The system performs self-adaptation through automated control operations that learn optimal settings from accumulated data.
Solution Approach 2:
The patent replaces manual mechanical adjustment with an automated electronic control system that uses sensors, processors, and reinforcement learning algorithms to determine optimal throttle valve settings, substituting human expertise with computational intelligence.
2Extent of automation
If reinforcement learning is used to automatically update relationship definition data, then manual effort is reduced, but the system complexity increases
Solution Approach 1:
The control device integrates multiple functions including state detection, reward calculation, relationship definition data storage, and automated adjustment operations within a single system. This multi-functionality manages complexity by consolidating components rather than adding separate systems.
Solution Approach 2:
The system implements reinforcement learning with feedback loops where the control device detects vehicle states, calculates rewards based on control performance, and uses this feedback to continuously update relationship definition data, creating a self-improving automated system.
3Reliability
If relationship definition data is updated based on single vehicle data, then the adaptation is vehicle-specific, but the update frequency is low
Solution Approach 1:
The control device combines and accumulates data from multiple vehicles to update the shared relationship definition data. By merging information across the vehicle fleet, the system achieves both vehicle-specific adaptation through aggregated patterns and high update frequency through pooled data collection.
Solution Approach 2:
The system performs preliminary data collection and accumulation across multiple vehicles before updating the relationship definition data. This preliminary action of gathering sufficient data from the fleet enables more frequent and reliable updates while maintaining vehicle-specific applicability.
Data Source
AI summary
A vehicle control system includes a memory, a first processor mounted in a vehicle, and a second processor different from an in-vehicle device. The first processor and the second processor are configured to execute acquisition processing, operation processing, reward calculation processing, and update processing. The first processor is configured to execute at least the acquisition processing and the operation processing, and the second processor is configured to execute the update processing.