Lane Region-Based Routing for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional routing systems for autonomous vehicles lack efficiency and flexibility in searching for optimal routes, failing to effectively handle lane changes and cover all possible paths.
Innovation Solution
A lane region-based or lane segment-based routing system is implemented, where lanes form a road-level topographic map to mark upstream, downstream, and neighboring lanes, allowing for efficient route decision and planning by identifying all possible routes from a starting point to a destination through lane regions and lane changes between adjacent regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional routing systems are used for autonomous vehicles, then the system structure is simple, but the routing efficiency and flexibility are insufficient
Solution Approach 1:
The patent segments the routing system into multiple hierarchical levels: road-level routing that determines sequences of roads, and lane-level routing that determines specific lanes within each road. This segmentation allows each level to operate independently with optimized algorithms, improving overall routing efficiency without requiring complete system redesign.
Solution Approach 2:
The patent introduces a two-dimensional routing framework where the first dimension is road-level sequencing and the second dimension is lane-level selection within each road. This dimensional approach enables efficient route planning by separating macro-level path determination from micro-level lane selection, thereby improving productivity without proportionally increasing complexity.
2Adaptability or versatility
If conventional routing systems are used, then the system is easy to implement, but the ability to handle lane changes and cover all possible paths is insufficient
Solution Approach 1:
The routing system is divided into road-level and lane-level components, where the road-level module determines the sequence of roads and the lane-level module handles specific lane selections and changes. This segmentation enables the system to adapt to various lane change scenarios without requiring complete reconfiguration of the entire routing system.
Solution Approach 2:
The patent implements dynamic lane-level routing that can adapt to real-time conditions, allowing the system to handle different lane change scenarios flexibly. The lane-level module can dynamically adjust lane selections based on traffic conditions, road geometry, and vehicle objectives, enhancing adaptability while maintaining manageable system complexity through modular design.
3Adaptability or versatility
If detailed lane-level routing is implemented, then routing flexibility improves, but computational resource consumption increases
Solution Approach 1:
By segmenting routing into road-level and lane-level tasks, the patent reduces computational burden. The road-level module performs coarse-grained path planning with lower computational requirements, while the lane-level module handles fine-grained lane selections only when necessary, thereby maintaining routing flexibility while reducing overall computational resource consumption.
Solution Approach 2:
The system applies lane-level detailed routing only when needed (e.g., at intersections, lane changes, or complex road segments) rather than continuously throughout the entire journey. This partial application of detailed routing maintains flexibility where required while minimizing computational resource consumption during straightforward driving segments.
Data Source
AI summary
In one embodiment, a routing request is received for routing an autonomous driving vehicle (ADV) from a source lane to a target lane. One or more road paths are determined from a source road to a target road. The road paths include zero or more intermediate roads in between, where each intermediate road includes one or more intermediate lanes. For each of the road paths, one or more lane paths are determined. Each lane path includes a number of lanes in combination to connect the source lane of the source road to the target lane of the target road via at least one of the intermediate lanes of the intermediate roads. A trajectory is planned from the source plane of the source road to the target lane of the target road using the lane paths to drive the ADV according to the trajectory.