Virtual Lane Estimation Using Neighbor Vehicle Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lane detection systems in autonomous vehicles rely on visual cues and often produce false positives, failing to account for unclear road markings or unexpected situations, and lack a model for reasonable driving behavior.
Innovation Solution
A virtual lane estimation system using a recursive self-organizing map that learns from the trajectories of neighboring vehicles to determine drivable lanes, generating a directed graph for navigation control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lane detection systems rely on visual cues from road markings, then lane detection can be performed using existing infrastructure, but the system produces false positive detections and fails when road markings are unclear or absent
Solution Approach 1:
The patent introduces an intermediary layer (virtual lane estimation system using recursive self-organizing maps) that mediates between raw sensor data and lane detection decisions. This intermediary processes trajectories of multiple vehicles to infer virtual lanes, reducing false positives from relying solely on visual road markings while maintaining system reliability.
Solution Approach 2:
The system creates virtual copies of lane information by tracking and analyzing trajectories of multiple neighboring vehicles. Instead of directly detecting road markings, the system copies motion patterns from other vehicles to infer drivable paths, enabling reliable lane detection even when visual cues are absent or ambiguous.
2Adaptability or versatility
If lane detection systems use traditional visual cue methods, then the system structure remains simple, but the system cannot account for unexpected situations or provide a model of reasonable driving behavior
Solution Approach 1:
The patent implements a dynamic system where the recursive self-organizing map continuously adapts to new trajectory data from neighboring vehicles. This dynamic learning capability enables the system to handle unexpected situations and evolving traffic patterns while maintaining a relatively simple architectural framework based on established neural network principles.
Solution Approach 2:
The virtual lane estimation system serves multiple functions: it detects lanes, models reasonable driving behavior, handles unexpected situations, and provides trajectory prediction. This multi-functional approach increases adaptability while avoiding the need for separate specialized systems for each function.
3Productivity
If the system trains the recursive self-organizing map using all trajectories simultaneously, then training convergence may be faster, but memory requirements and computational complexity increase significantly
Solution Approach 1:
The patent segments the training process by processing trajectories sequentially or in small batches rather than loading all trajectory data simultaneously into memory. This segmentation approach maintains training productivity while significantly reducing memory requirements, allowing the system to handle large datasets on resource-constrained autonomous vehicle platforms.
Data Source
AI summary
A virtual lane estimation system includes a memory device, a sensor and a computer. The memory device is configured to store a road map that corresponds to a portion of a road ahead of a vehicle. The sensor is configured to observe a plurality of trajectories of a plurality of neighboring vehicles that traverse the portion of the road. The computer is configured to initialize a recursive self-organizing map as a plurality of points arranged as a two-dimensional grid aligned with the road map, train the points in the recursive self-organizing map in response to the trajectories, generate a directed graph that contains one or more virtual lanes through the road map in response to the points trained to the trajectories, and generate a control signal that controls navigation of the vehicle through the portion of the road in response to the virtual lanes in the directed graph.