Autonomous Vehicle Speed Control Using Human Driving Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles struggle to interact safely with human-driven vehicles and environments due to the complexity of human driving behaviors, requiring systems that can mimic human-like driving decisions based on various environmental factors.
Innovation Solution
A machine-learned model, such as a neural network, trained on human driving data, determines vehicle speed by analyzing driving environment characteristics like lane width, traffic speed, and visibility, allowing the vehicle to adjust speed in a manner similar to human drivers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a machine-learned model trained on human driving data is used to determine target vehicle speed, then the autonomous vehicle's ability to interact safely and effectively with human-driven vehicles improves, but the system complexity increases
Solution Approach 1:
The patent replaces traditional physics-based control models with a machine-learned model (neural network) that processes sensor data and directly outputs target speed commands. This substitution allows the system to capture complex human driving behaviors and environmental interactions that cannot be modeled through physical equations alone, thereby improving safety in human-vehicle interactions while accepting increased computational complexity
Solution Approach 2:
The machine-learned model is trained on driving data collected from human operators, effectively copying human driving patterns and decision-making processes. This copying approach enables the autonomous vehicle to mimic human-like speed selection in various driving scenarios, improving its ability to interact predictably and safely with human-driven vehicles
2Measurement precision
If a machine-learned model is used to determine target vehicle speed based on driving environment characteristics, then the accuracy and contextual appropriateness of speed decisions improves, but the computational processing requirements increase
Solution Approach 1:
The machine-learned model is pre-trained offline on extensive driving data collected from human operators across diverse conditions. This preliminary training phase allows the model to learn complex patterns and relationships before deployment, enabling accurate speed decisions during actual operation without requiring intensive real-time computation, thus balancing accuracy with energy efficiency
Data Source
AI summary
A machine-learned model is trained using human driving data to determine a desired vehicle speed based from a set of driving-environment characteristics. An autonomous-vehicle control system obtains, from cameras, sensors, services, and data sources, a variety of sensor data. The sensor data is used to determine a set of characteristics for the driving-environment for the autonomous vehicle. Using the machine-learned model, the autonomous-vehicle control system determines a human-like desired speed for the autonomous vehicle based at least in part on the determined characteristics of the driving-environment.