Vehicle Sensor Anomaly Detection for Reliable Map Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems lack comprehensive data gathering methods to accurately detect and update navigation routes in response to anomalous driving conditions, particularly for autonomous vehicles, as they rely heavily on user feedback which may be incomplete.
Innovation Solution
A system that utilizes non-geospatial vehicle sensor data, such as speedometer and steering wheel position data, to detect anomalous driving conditions through a trained model, and updates maps when a threshold of vehicles report similar conditions, enabling proactive route adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If user feedback is used to gather traffic data, then direct user input can be obtained, but the data coverage is incomplete as users may not frequent all road segments
Solution Approach 1:
The system enables vehicles to automatically report anomalous driving conditions through sensor data without requiring manual user input. The trained model detects anomalies from vehicle sensor data and automatically updates map segments, allowing the system to self-gather traffic information while maintaining data coverage.
Solution Approach 2:
The patent replaces manual user feedback mechanisms with automated sensor-based detection systems. Vehicle sensors continuously monitor driving conditions and feed data to the trained model, which automatically identifies anomalies and triggers map updates without requiring user interaction.
2Measurement precision
If a trained model is used to detect anomalous driving conditions, then detection accuracy is improved, but system complexity increases
Solution Approach 1:
The trained model serves as an intermediary between raw vehicle sensor data and map update decisions. It processes sensor data to detect anomalous driving conditions, filtering and interpreting the data before triggering map segment updates, thereby improving detection accuracy while managing system complexity through modular architecture.
3Reliability
If map segments are marked based on single vehicle data, then quick updates are possible, but reliability decreases due to potential false positives
Solution Approach 1:
The system performs preliminary comparison of anomalous driving condition data across multiple vehicles before marking map segments. By pre-comparing data from multiple sources and applying threshold criteria, the system ensures reliable map updates only when multiple vehicles confirm the same anomaly, reducing false positives while maintaining efficient update timing.
Data Source
AI summary
Systems, methods and vehicles for detecting anomalous driving conditions are disclosed. In one embodiment, a method includes receiving vehicle sensor data from a vehicle, where a portion of the vehicle sensor data is non-geospatial data, inputting the vehicle sensor data into a trained model that is trained to detect anomalous driving conditions, when an output of the trained model indicates an anomalous driving condition, comparing the vehicle sensor data associated with the anomalous driving condition with vehicle sensor data of one or more additional vehicles according to a metric, and when the metric is satisfied, marking a segment of a map corresponding with the vehicle data associated with the anomalous driving condition, and transmitting map data associated with the map to the vehicle.