Vehicle Path Prediction Using Future Frames for Occluded Lanes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current path prediction systems for intelligent driving monitoring systems (IDMS) and autonomous driving systems face challenges in situations with unclear or occluded lane markings, requiring robust methods to determine a safe path of travel, especially in complex and dynamic real-world scenarios.
Innovation Solution
The 'Back-to-the-Future' approach uses future frames and sensor data to determine a path of travel, incorporating information from 'future' data relative to a given time, enabling robust path prediction even in challenging conditions by mapping the path to a camera reference frame and using a combination of sensors like GPS, Lidar, and radar, and machine learning models for improved accuracy and reduced latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional path prediction methods are used, then the system can operate with current data only, but the accuracy and reliability deteriorate in situations with unclear or occluded lane markings
Solution Approach 1:
The system performs preliminary actions by collecting and storing future frame data and sensor information in advance. This allows the path prediction system to access previously captured future states when current lane markings are unclear or occluded, improving reliability without requiring complex real-time processing during critical moments.
Solution Approach 2:
The system introduces an intermediary mechanism by mapping determined paths from future frames back to the current camera reference frame. This intermediary mapping process allows the system to bridge the gap between historical future data and current decision-making, enabling reliable path prediction even when direct current observations are insufficient.
2Measurement precision
If future frames and sensor data are incorporated to improve path prediction accuracy, then the reliability improves, but the processing complexity and data handling requirements increase
Solution Approach 1:
The system extracts only the essential path information from future frames and sensor data, rather than processing all raw data. By extracting and storing only the determined paths and relevant features for later mapping, the system achieves high prediction accuracy while minimizing the complexity of data handling and processing requirements.
3Loss of time
If real-time or near-real-time path prediction is implemented, then the response time is reduced, but the accuracy may deteriorate due to limited processing time
Solution Approach 1:
The system performs preliminary path determination on future frames in advance, storing the results for later retrieval and mapping. This preliminary processing allows the system to provide real-time or near-real-time predictions with high accuracy, as the computationally intensive path analysis has already been completed before the critical decision moment arrives.
Data Source
Figure 1
Figure 2
Figure 3A~3C
AI summary
The present disclosure provides systems and methods for mapping a determined path of travel. The path of travel may be mapped to a camera view of a camera affixed to a vehicle. In some embodiments, the path of travel may be mapped to another view that is based on a camera, such as a bird's eye view anchored to the camera's position at a given time. These systems and methods may determine the path of travel by incorporating data from later points in time.