Neighboring Vehicle Path Prediction Using Detailed Map Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle collision warning systems, such as adaptive cruise control and forward collision warning, face challenges in accurately predicting the driving path of neighboring vehicles in complex environments due to limitations in radar sensors, leading to misrecognition and decreased reliability.
Innovation Solution
A method and system that utilize a detailed map to detect segments adjacent to neighboring vehicles, employing grid maps or quad tree methods to predict their driving paths using velocity, acceleration, and driving direction, and providing options for most probable, probability estimation, and defensive estimation paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If radar sensors are used to detect neighboring vehicles, then the system can operate regardless of weather conditions, but the measurement precision and reliability decrease in complex driving environments
Solution Approach 1:
The patent combines multiple sensing technologies (radar, camera, LIDAR, ultrasonic sensors) into a fused sensing system that leverages the strengths of each sensor type while compensating for their individual weaknesses, thereby maintaining measurement precision across diverse driving conditions
Solution Approach 2:
The patent introduces map information and road structure data as intermediary elements that mediate between sensor inputs and vehicle path prediction, using pre-known road geometry to disambiguate sensor measurements and improve detection accuracy in complex environments
2Speed
If radar is used for forward collision warning, then the system can detect vehicles at a distance, but it cannot distinguish multiple adjacent vehicles or recognize stopped vehicles due to diffused reflection
Solution Approach 1:
The patent segments the detection space into multiple zones and uses multiple sensing devices positioned at different locations to capture data from different segments, enabling the system to distinguish between multiple vehicles that would appear as a single target to a single radar sensor
Solution Approach 2:
The patent creates a multi-functional sensing system where different sensor types serve multiple purposes: radar provides long-range detection and velocity measurement, cameras provide visual identification and stationary object detection, and LIDAR provides precise distance measurement, with each sensor type compensating for the limitations of others
3Speed
If sensor data is used to predict driving direction, then the system can provide real-time warnings, but misrecognition occurs leading to decreased reliability
Solution Approach 1:
The patent implements feedback mechanisms where the system continuously monitors sensor data, compares predicted paths with actual vehicle movements, and adjusts prediction algorithms based on discrepancies, thereby improving reliability while maintaining real-time response capability
Solution Approach 2:
The patent performs preliminary path prediction using current sensor data and map information before actual collision risk materializes, allowing the system to prepare warning signals and navigation adjustments in advance while maintaining high reliability through pre-computed safe paths
Data Source
AI summary
A method and system for predicting a driving path of a neighboring vehicle which may influence a subject vehicle is provided. The subject vehicle detects a segment adjacent to the neighboring vehicle which is currently driven and searches all segments connected to the segment. The method prevents a collision between the subject vehicle and the neighboring vehicle that occurs due to misrecognition of a position, a driving direction, or velocity of the neighboring vehicle, and more accurately predicts the driving path of the neighboring vehicle. The method includes receiving information of a neighboring road on which a subject vehicle is driven from a detailed map and detecting a segment showing the neighboring vehicle which is being currently driven from the information of the neighboring road. Segments are detected that are connected to the segment and the driving path of the neighboring vehicle is predicted.