Autonomous Vehicle Controller Learning Driver Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicles have limitations in reflecting the driving patterns of various drivers, leading to a sense of difference between real and autonomous driving, and struggle to adapt to unpredictable road environments, resulting in passive and inflexible responses that compromise ride comfort and safety.
Innovation Solution
An apparatus and method for controlling autonomous vehicles that includes a user input unit, information collection unit, and controller to learn and adapt to a driver's driving pattern by identifying learning sections, updating autonomous driving control values based on driver operation information, and switching between autonomous and manual driving modes to enhance ride comfort and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If autonomous vehicles use a monolithic driving strategy with fixed parameters, then safety is ensured through passive correspondence, but the system cannot adapt to unpredictable road environments and various driver tendencies
Solution Approach 1:
The patent divides the driving environment into multiple learning sections (e.g., congestion sections, acceleration sections, deceleration sections, curved sections, downhill sections, toll collection sections) along the route. Each section can be independently learned and adapted, allowing the system to handle complex environments through modular segmentation rather than requiring a completely complex monolithic system.
Solution Approach 2:
The system dynamically switches between autonomous driving mode and manual driving mode based on the current learning section. During manual mode, the system collects driver operation information to update driving patterns. This dynamic adaptability allows the system to respond flexibly to changing environments without permanently increasing system complexity.
2Ease of operation
If autonomous vehicles follow a single monolithic driving strategy, then control is simplified, but the system creates a sense of difference between real driving and autonomous driving, reducing ride comfort
Solution Approach 1:
The system collects driver operation information (steering wheel operations, accelerator pedal operations, brake pedal operations) during manual driving mode and uses this feedback to learn and update the driving pattern. This feedback mechanism allows the system to gradually adapt to individual driver tendencies, making autonomous driving more comfortable and natural for each driver.
Solution Approach 2:
The system learns and updates driving pattern parameters by comparing autonomous driving control values with actual driver operations. By adjusting parameters such as steering angle, acceleration, and deceleration based on learned patterns, the system can mimic individual driver styles, improving ride comfort while maintaining adaptability.
3Reliability
If autonomous vehicles perform passive correspondence focused on safety, then collision prevention is achieved, but the system lacks flexible response to unpredictable situations
Solution Approach 1:
The system performs preliminary learning by collecting driver operation information in advance during manual driving mode. This pre-collected data is stored and used to predict appropriate driver responses in similar future situations, enabling flexible correspondence while maintaining safety through proven driver-like decision-making patterns.
Solution Approach 2:
The system serves itself by automatically learning and updating its own driving patterns without external intervention. Through continuous learning from driver operations, the system improves its ability to handle unpredictable situations flexibly while maintaining the safety foundation established by its autonomous driving algorithms.
Data Source
AI summary
An apparatus and method for controlling an autonomous vehicle are provided. The apparatus includes a user input unit that receives identification information of a driver within the vehicle during autonomous driving and an information collection unit that acquires a global route of the vehicle and surrounding environment information. A controller determines a learning section on the global route based on the surrounding environment information and outputs a driving pattern of the driver by performing repetitive learning based on operation information of the driver in the learning section. Autonomous driving of the vehicle is then executed based on the output driving pattern of the driver.