Autonomous Vehicle Positioning Using Lane-Level Trip Demand
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining optimal deployment locations for autonomous vehicles within a city to maximize usage and profitability, while minimizing intervention events, is challenging due to the need for accurate mapping and identification of suitable road networks and pickup/drop-off zones.
Innovation Solution
A system that utilizes historical trip data to evaluate lane-level performance metrics, identify high-potential pickup and drop-off zones, and optimize vehicle deployment by processing geohash data and implementing a recommender algorithm to prioritize road networks with low intervention probability and high trip counts, ensuring efficient AV operation and profitability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If autonomous vehicles are deployed to maximize trip demand, then profitability is improved, but intervention events increase
Solution Approach 1:
The system performs preliminary analysis of historical trip data and road network characteristics before deploying autonomous vehicles. By pre-identifying high-demand locations and routes with low intervention probability, the system optimizes vehicle placement in advance to maximize trip demand while minimizing intervention events, resolving the contradiction between productivity and reliability.
2Adaptability or versatility
If comprehensive map data is collected for entire city, then AV travel coverage is improved, but data processing complexity increases
Solution Approach 1:
The system extracts and focuses only on the most relevant features from comprehensive map data, such as road network topology, traffic patterns, and intervention-prone areas. By taking out only the essential data elements needed for AV navigation and deployment optimization, the system maintains broad travel coverage while reducing data processing complexity.
Solution Approach 2:
The system segments the city map into distinct zones or regions based on traffic characteristics, demand patterns, and intervention probabilities. This segmentation allows the AV system to process and manage map data in smaller, more manageable units, reducing overall complexity while maintaining comprehensive coverage across the entire city.
3Productivity
If AVs are positioned to maximize usage, then profitability is improved, but mapping requirements increase
Solution Approach 1:
The system performs preliminary identification of optimal vehicle positioning locations by analyzing historical trip data and road network characteristics before deployment. By pre-determining high-utilization zones and their corresponding mapping requirements, the system optimizes vehicle placement to maximize usage while minimizing the extent of mapping needed.
Data Source
AI summary
Techniques are provided for determining where to position vehicles for trip optimization or where to map roads for use by autonomous or semi-autonomous vehicles. The techniques include identifying, from historical trip data, common pickup and drop-off points within a geographical area where respective geohashes are used as nodes in the geographical area. A number of trips between respective nodes in the geographical area within a predetermined time frame define edges between respective nodes in the geographic area. The nodes and edges for the geographic area are processed to score each node to identify most active nodes within the geographic area as potential pickup/drop-off zones. The top k potential pickup-drop-off zones are evaluated for suitability as a pickup/drop-off zone, and lane IDs, suitable pickup/drop-off zones, and/or trip lists derived from the historical trip data are provided for use in positioning vehicles or mapping roads.