Autonomous Vehicle Routing Using Spatiotemporal Event Risk
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional autonomous vehicle routing systems rely on predefined metrics such as travel time and distance, which may result in a suboptimal experience due to the lack of consideration for operation-influencing events and spatiotemporal factors, potentially leading to undesirable driving scenarios.
Innovation Solution
A computer-implemented spatiotemporal statistical model is used to generate a score indicating the likelihood of operation-influencing events based on historical data, allowing the autonomous vehicle to select a route that balances travel time with the risk of encountering such events, thereby optimizing the driving experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional predefined metrics (travel time, distance) are used for routing, then routing simplicity is maintained, but driving experience quality deteriorates due to lack of consideration for operation-influencing events
Solution Approach 1:
The system performs preliminary actions by pre-processing historical operational data to train machine learning models that predict operation-influencing events. These models are trained offline using historical data about events such as human operator interventions, unplanned maneuvers, and deceleration events, enabling the routing system to anticipate potential issues before they occur without adding real-time computational complexity
Solution Approach 2:
The patent introduces an intermediary layer between conventional routing and the autonomous vehicle operation. This intermediary consists of machine learning models that process spatiotemporal features (weather, traffic, road conditions) and historical event data to generate predictions about operation-influencing events. These predictions are then integrated into the routing decision-making process, allowing the system to balance simplicity with improved reliability
2Productivity
If route selection considers only travel time and distance, then routing efficiency is maintained, but safety deteriorates due to exposure to challenging driving scenarios
Solution Approach 1:
The system performs preliminary risk assessment by training machine learning models on historical data about operation-influencing events before route selection. These models predict the likelihood of safety-related events (human takeovers, unplanned maneuvers, excessive deceleration) for different routes based on spatiotemporal features, allowing the system to avoid high-risk areas proactively while maintaining routing efficiency
Solution Approach 2:
The system implements feedback mechanisms where historical operational data from autonomous vehicle events is continuously collected and used to retrain and improve the machine learning models. This feedback loop allows the system to learn from past safety incidents and improve its ability to predict and avoid challenging driving scenarios, progressively enhancing safety performance
3Reliability
If spatiotemporal factors are incorporated into routing decisions, then driving experience is improved, but system complexity increases due to additional data processing requirements
Solution Approach 1:
The patent segments the complex routing problem into distinct components: spatiotemporal feature extraction (weather, traffic, road conditions), historical event data processing, machine learning model predictions, and route optimization. Each component is handled by specialized modules that can be independently trained and optimized, reducing overall system complexity while maintaining comprehensive analysis capabilities
Solution Approach 2:
The system performs preliminary processing of large volumes of historical data and training of machine learning models offline before deployment. This pre-processing reduces the computational burden during real-time route selection, as the models have already learned patterns from historical data. The runtime system only needs to query pre-trained models with current spatiotemporal features, significantly reducing real-time complexity
Data Source
AI summary
Various technologies described herein pertain to routing autonomous vehicles based upon spatiotemporal factors. A computing system receives an origin location and a destination location of an autonomous vehicle. The computing system identifies a route for the autonomous vehicle to follow from the origin location to the destination location based upon output of a spatiotemporal statistical model. The spatiotemporal statistical model is generated based upon historical data from autonomous vehicles when the autonomous vehicles undergo operation-influencing events. The spatiotemporal statistical model takes, as input, a location, a time, and a direction of travel of the autonomous vehicle. The spatiotemporal statistical model outputs a score that is indicative of a likelihood that the autonomous vehicle will undergo an operation-influencing event due to the autonomous vehicle encountering a spatiotemporal factor along a candidate route. The autonomous vehicle then follows the route from the origin location to the destination location.