Vehicle Driving Control Using Deep Learning for Friction Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional autonomous driving vehicles struggle to accurately analyze the influence of factors such as air resistance and friction during travel, which hinders effective driving control.
Innovation Solution
An information processing device utilizing deep learning for multivariate analysis to infer index values from various vehicle-related data, enabling real-time driving control with high accuracy by integrating sensor inputs and performing calculations at nanosecond intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional autonomous driving vehicles use traditional control methods, then device complexity is reduced, but measurement precision of environmental factors such as air resistance and friction deteriorates
Solution Approach 1:
The patent replaces traditional mechanical control systems with an information processing device that uses deep learning algorithms to analyze environmental factors. The inference unit processes sensor data through neural networks to calculate air resistance, friction, and other environmental influences, achieving high measurement precision without mechanical measurement devices.
Solution Approach 2:
The patent introduces an information processing device as an intermediary between sensors and driving control. This intermediary unit processes raw sensor data through deep learning models to extract meaningful environmental factor values, enabling accurate analysis while maintaining system modularity and managing complexity.
2Measurement precision
If deep learning multivariate analysis is implemented, then driving control accuracy is improved, but calculation time increases
Solution Approach 1:
The patent performs preliminary actions by pre-training deep learning models with extensive environmental data before actual driving operations. The inference unit uses these pre-trained models to quickly infer environmental factors during real-time operation, reducing calculation time while maintaining high accuracy through previously learned patterns.
Solution Approach 2:
The patent changes parameters by optimizing deep learning model architecture and inference settings to balance accuracy and speed. The system adjusts model complexity, input data selection, and calculation precision dynamically to achieve optimal driving control accuracy within acceptable time constraints.
3Speed
If real-time multivariate analysis is performed at nanosecond intervals, then driving control responsiveness is improved, but energy consumption increases
Solution Approach 1:
The patent implements a universal information processing device that handles multiple functions: sensor data acquisition, deep learning inference, environmental factor calculation, and driving control commands. This multi-functional integration reduces overall system energy consumption by sharing processing resources while maintaining nanosecond-level responsiveness through optimized architecture.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An information processing device of the disclosure includes an information acquisition unit capable of acquiring plural pieces of information related to a vehicle, an inference unit that uses deep learning to infer plural index values from the plural pieces of information acquired by the information acquisition unit, and a driving control unit that executes driving control of the vehicle on the basis of the plural index values.