Driver Risk Grouping With Selective Alerts for Fleet Coaching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver monitoring systems struggle to effectively detect and address unsafe and inefficient driving behaviors, particularly those not accompanied by large inertial sensor readings, leading to overwhelmed safety management and coaching resources in fleets.
Innovation Solution
A method and system utilizing camera sensors and inertial sensors to detect traffic events, coupled with analytical methods to calculate a driver score based on detected driving events, allowing for the allocation of drivers into coachable risk groups and selective transmission of alerts to improve driving behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If video processing-based systems are used to detect driving events, then the number and range of detected driving events increases, but the available data overwhelms safety management and coaching resources
Solution Approach 1:
The patent segments the continuous video data stream into discrete, analyzable driving events by detecting specific patterns and behaviors. This segmentation transforms overwhelming continuous data into manageable discrete events that can be individually assessed and coached, resolving the contradiction between comprehensive detection and manageable data volume.
Solution Approach 2:
The patent introduces an intermediary processing layer that filters and prioritizes video data before presenting it to safety management systems. This intermediary layer processes the raw video feed, identifies significant driving events, and presents a curated set of events for coaching, thereby managing the data flow between detection and coaching resources.
2Measurement precision
If inertial sensor readings are used to detect unsafe driving events, then large inertial readings can be detected, but certain common driving events without large inertial readings are missed
Solution Approach 1:
The patent merges inertial sensor data with video processing data to create a comprehensive detection system. By combining these two data sources, the system benefits from the precision of inertial sensors for high-impact events while gaining the versatility of video processing for detecting subtle driving behaviors that inertial sensors alone would miss.
Solution Approach 2:
The patent creates a multi-functional detection system that can handle both high-impact events (detected by inertial sensors) and subtle driving behaviors (detected by video processing). This universal system adapts to different types of driving events, providing comprehensive coverage across the full range of unsafe driving behaviors.
3Reliability
If all detected driving events are reviewed by human operators, then comprehensive safety assessment is achieved, but safety management and coaching resources become overwhelmed
Solution Approach 1:
The patent applies partial action by having human operators review only a subset of detected driving events - specifically those that are most significant or ambiguous. The system automatically processes and filters events, presenting only those requiring human judgment, thereby maintaining safety assessment quality while improving coaching efficiency.
Solution Approach 2:
The patent implements feedback mechanisms where the system learns from human operator reviews of driving events. This feedback loop allows the system to improve its automatic filtering and prioritization capabilities over time, reducing the burden on human operators while maintaining or improving safety assessment quality.
Data Source
AI summary
Systems and methods for selectively transmitting alerts based on monitored behavior of a driver are disclosed. A system can calculate a driver score for a driver based on a set of driving events detected from sensor data captured while a vehicle was operated by the driver. Each of the set of driving events is associated with a driving behavior in a driving scenario. The driver score is further calculated based on historic driving event data. The system can assign the driver to a category of a number of categories based on the driver score. The system can determine that a change to a habitual driving behavior in a monitored driving scenario would result in reassignment of the driver to another category. The system can transmit an alert to the vehicle based upon a subsequent detection of a driving event that is associated with the monitored driving scenario.