Feature Database for Autonomous Vehicle Route Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating through environments with dynamic and static features, such as traffic controls and road conditions, which affect route planning and safety, as existing systems lack detailed understanding of features contributing to travel times.
Innovation Solution
A feature database system that updates and classifies semantic features like traffic lights, road signs, and road damage using sensor data, allowing for accurate route planning by determining feature costs and impacts on travel time, safety, and desirability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If autonomous vehicles use traditional route planning methods without detailed feature data, then the system complexity is reduced, but the accuracy of route planning and understanding of travel time impacts deteriorates
Solution Approach 1:
The system performs preliminary detection and classification of semantic features (traffic lights, stop signs, road conditions) and stores them in a feature database before route planning occurs. This pre-processing of environmental data allows accurate route planning without increasing real-time system complexity, as the feature information is already organized and ready for query during route determination.
Solution Approach 2:
A feature database acts as an intermediary between sensor data and route planning algorithms. The database stores classified semantic features and their associated travel time costs, decoupling the complexity of feature detection from route planning. This intermediary layer allows the route planning system to query pre-processed feature information without directly handling the complexity of real-time sensor processing.
2Loss of information
If autonomous vehicles collect and process detailed sensor data about environmental features, then the understanding of feature impacts on travel time improves, but the loss of time for data processing increases
Solution Approach 1:
The system detects, classifies, and stores semantic features in the feature database in advance, before they are needed for route planning. By performing this data processing preliminarily and organizing features with their associated travel time costs in the database, the system avoids time-consuming processing during actual route planning, thus reducing information loss without proportionally increasing processing time loss.
Solution Approach 2:
The feature database automatically updates itself with new semantic feature detections from sensor data without requiring manual intervention. The system self-maintains the feature information repository, classifying features and determining their travel time impacts autonomously. This self-service approach minimizes the time loss associated with data processing by making the system self-sufficient in maintaining accurate feature information.
3Productivity
If autonomous vehicles rely on manual updates of route information, then the system complexity is reduced, but the productivity of route planning deteriorates
Solution Approach 1:
The feature database automatically detects new semantic features from sensor data, classifies them, determines their travel time costs, and updates the database without manual intervention. This self-service automation significantly improves route planning productivity by continuously maintaining up-to-date feature information. The increased automation complexity is offset by the substantial gains in route planning efficiency and the elimination of manual data updates.
Data Source
AI summary
Techniques for determining a location and type of traffic-related features and using such features in route planning are discussed herein. The techniques may include determining that sensor data represents a feature such as a traffic light, road segment, building type, and the like. The techniques further include determining a cost of a feature, whereby the cost is associated with the effect of a feature on a vehicle traversing an environment. A feature database can be updated based on features in the environment. A cost of the feature can be used to update costs of routes associated with the location of the feature.