Blind Spot Course Prediction for Curved Road Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driving support systems in vehicles fail to accurately predict the course of objects in blind spots due to static sensor configurations and linear estimation methods, leading to potential errors and stopped movement prediction when objects enter these areas.
Innovation Solution
A course prediction device with a blind spot calculation unit and prediction unit that dynamically calculates the blind spot area using sensor information and predicts the course of objects within this area, accounting for complex road shapes and multiple objects in the depth direction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If linear estimation method is used to predict object position in blind spot area, then calculation simplicity is improved, but position estimation accuracy deteriorates in complicated road shapes
Solution Approach 1:
The patent transitions from static linear estimation to dynamic curve fitting prediction. The prediction unit dynamically determines prediction curves based on detected object movement patterns, adapting the prediction model to match actual object trajectories in complex road environments, thereby improving accuracy while maintaining computational efficiency
Solution Approach 2:
The patent changes the fundamental parameter of prediction from linear to curved trajectories. By using second-degree or higher polynomial curves instead of straight lines, the system can accurately represent objects moving along curved roads or changing direction, significantly improving position estimation accuracy in complicated road shapes
2Speed
If static blind spot area is obtained from sensor specifications, then calculation speed is improved, but detection capability deteriorates when multiple objects are lined up in depth direction
Solution Approach 1:
The patent implements dynamic blind spot area calculation that updates in real-time based on detected object positions. Instead of using a fixed static blind spot region from sensor specifications, the system continuously recalculates the blind spot area considering the latest object detection data, enabling accurate tracking of multiple objects lined up in the depth direction while maintaining real-time performance
Solution Approach 2:
The system uses feedback from sensor detection results to continuously update and refine the blind spot area calculation. By incorporating actual detection data into the blind spot determination process, the system adapts to changing environmental conditions and object configurations, improving detection reliability for multiple objects in depth without sacrificing calculation speed
3Device complexity
If movement prediction is stopped when object enters blind spot area, then system simplicity is improved, but driving support reliability deteriorates due to erroneous behavior assumption
Solution Approach 1:
The patent applies preliminary action by predicting object courses in advance before they enter or while they are in the blind spot area. The prediction unit continuously calculates future object positions using curve fitting methods, ensuring that movement prediction information is available even when objects are in undetectable regions, thereby maintaining driving support reliability without complicating the system architecture
Solution Approach 2:
The prediction unit acts as an intermediary that bridges the gap between sensor detection limitations and driving support requirements. By generating predicted position data for objects in blind spots, the prediction unit provides continuous movement information to the driving support system, preventing erroneous behavior assumptions while keeping the overall system structure simple and manageable
Data Source
AI summary
A course prediction device (1) includes a blind spot calculation unit (11) and a prediction unit (12). The blind spot calculation unit (11) acquires a sensing result sequentially from a sensor which is arranged in a mobile object and which detects whether an object exists, and calculates, based on sensor information, a blind spot area expressing such an area that an object existing therein cannot be detected by the sensor. The blind spot calculation unit (11) also detects an object entering the blind spot area based on the sensor information. The prediction unit (12) predicts a course in the blind spot area of the detected object based on the sensing result.