Highway Egress Route Graph for Incomplete Lane Map Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
At least partially autonomous vehicles face challenges in determining an egress route on at least partially controlled-access multilane highways due to incomplete or missing map data, particularly when approaching a target off-ramp, requiring driver intervention.
Innovation Solution
A method and automotive control unit that generate an egress route graph based on available map data, assigning confidence and cost values to edges, and determine an optimized route to a target off-ramp using Dijkstra's algorithm, while updating the route based on sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If map data is used to determine egress route on controlled-access highway, then route determination is enabled, but incomplete or missing map data causes inability to determine longitudinal and lateral control
Solution Approach 1:
The system performs preliminary actions by detecting lane endings ahead of time using sensor data and map data integration. When a lane is detected to end, the system proactively determines alternative routes and prepares egress maneuvers before the lane actually ends, ensuring continuous autonomous control even when map data is incomplete.
Solution Approach 2:
Sensor data acts as an intermediary to compensate for incomplete map data. The system integrates sensor detections (lane markings, road geometry) with available map data to infer missing lane relationship information, enabling route determination without complete pre-mapped lane connectivity data.
2Ease of operation
If driver takes over control when map data is incomplete, then vehicle control is maintained, but autonomous operation is interrupted and driver intervention is required
Solution Approach 1:
The autonomous vehicle performs self-service by automatically detecting lane endings, determining alternative egress routes, and executing lane changes without driver intervention. The system uses its own sensor data to supplement map data gaps and autonomously resolves routing issues that would otherwise require driver input.
3Productivity
If traditional route optimization minimizes lane changes, then route efficiency is improved, but it cannot handle sections with missing lane relationship information
Solution Approach 1:
The route optimization is made dynamic by continuously updating the egress route graph with real-time sensor data. When map data is incomplete, the system dynamically infers lane relationships from sensor detections and adjusts the optimal route accordingly, rather than relying on static pre-computed routes that assume complete map data.
Data Source
Figure 1A~1C
Figure 2
Figure 3
AI summary
The present application relates to a method for determining an egress route for a vehicle to a target off-ramp of a highway, as well as a corresponding automotive control unit and a vehicle. To determine the egress route, an egress route graph is generated based on map data arranged in links. The egress route graph comprises a plurality of egress route nodes and a plurality of egress route edges, which includes lane direction edges based on lane information included in at least some of the links and lane change edges. To each egress route edge, a confidence value and a cost value are assigned. Based on the accordingly determined egress route graph, the egress route is determined, which corresponds to a path through the egress route graph to the target off-ramp, which has an optimized total confidence value and an optimized total cost value.