Driving Assistance Deceleration Learning for Recurring Brake Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving assistance systems fail to accurately determine whether a deceleration operation is a one-time event or a recurring requirement, leading to unnecessary learning of deceleration positions and prolonged completion time, and may erroneously execute deceleration control at one-time deceleration positions.
Innovation Solution
A driving assistance device that determines deceleration operations based on environmental factors ahead of the vehicle, assigning higher scores to positions where deceleration is likely to be recurring, such as stop lines or curves, and executes deceleration control only when a threshold score is reached, thereby reducing erroneous executions and accelerating learning completion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a predetermined number of deceleration operation records are required to prevent learning one-time deceleration positions, then the possibility of erroneous deceleration control is reduced, but the period required until learning of deceleration positions is completed increases
Solution Approach 1:
The patent applies local quality by assigning different score values to different deceleration positions based on their characteristics. Positions with deceleration request objects (stop lines, curves, intersections) receive higher scores, while ordinary positions receive lower scores. This differentiated scoring approach allows the system to quickly identify and learn important deceleration positions without requiring multiple passes, thereby reducing the learning completion period while maintaining high accuracy.
Solution Approach 2:
The patent changes the parameter of score assignment from uniform to variable based on environmental context. By introducing a scoring mechanism that considers the presence of deceleration request objects and road geometry, the system transforms the learning process from a quantity-based approach (number of records) to a quality-based approach (score threshold), significantly accelerating learning completion while preventing erroneous control at one-time deceleration positions.
2Measurement precision
If the vehicle travels on a new route many times before learning deceleration positions is completed, then the learning accuracy is improved, but the productivity and efficiency of the system deteriorates
Solution Approach 1:
The patent implements preliminary action by proactively identifying and scoring deceleration positions during the first pass through a route. When the vehicle encounters a deceleration request object or specific road geometry, the system immediately assigns a high score to that position, preparing the information in advance. This eliminates the need for multiple traversal cycles to learn deceleration positions, thereby maintaining high learning accuracy while significantly improving system productivity and efficiency.
Solution Approach 2:
The system uses feedback from environmental sensors (camera, radar, map information) to continuously evaluate and update the scores of deceleration positions. This real-time feedback mechanism allows the system to accurately identify recurring deceleration points on new routes during the first encounter, achieving high learning accuracy without requiring multiple passes, thus maintaining high system efficiency.
Data Source
AI summary
When the vehicle reaches the deceleration position where the driver performs the deceleration operation, the driving assistance device executes deceleration control for automatically decelerating the vehicle. The driving assistance device adds a larger score to the learning score of the deceleration position than the case where it is determined that the deceleration operation is performed due to the deceleration request object provided in front of the deceleration position and the vehicle needs to decelerate, and the deceleration operation is not performed due to the deceleration request object, on the basis of the environment in front of the vehicle when the driver performs the deceleration operation, and executes the deceleration control when the learning score of the deceleration position reached by the vehicle is equal to or larger than the threshold value.