Autonomous Vehicle Traction Prediction From Observed Object Acceleration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for measuring on-road traction in autonomous vehicles are limited to providing information about the road surface directly beneath the vehicle and cannot accurately estimate traction for surfaces in the forward path, relying on indirect methods and rules of thumb that may not be accurate in all driving environments.
Innovation Solution
An autonomous vehicle (AV) system that uses sensor data to compute the acceleration of objects in its driving environment, allowing it to determine the traction of the road surface ahead by observing the acceleration of objects traversing that surface, and shares this data with a fleet of AVs for coordinated maneuvering and route planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional sensor systems are used to measure traction, then the system can only provide information about the road surface directly beneath the vehicle, but it cannot accurately estimate traction for surfaces in the forward path
Solution Approach 1:
The system performs preliminary observation of objects traversing the forward path to detect their acceleration characteristics before the autonomous vehicle reaches that area. By analyzing how objects accelerate on upcoming road surfaces, the system predicts traction conditions in advance, allowing the vehicle to prepare appropriate control actions before actually encountering the surface.
2Ease of operation
If indirect methods such as temperature or windshield wiper duty cycle are used to estimate traction, then the system can provide traction estimates, but the estimates are inaccurate and rely on rules of thumb that may not be accurate in all driving environments
Solution Approach 1:
The system replaces indirect environmental inference methods with direct physical observation of object acceleration. Instead of using rules of thumb based on temperature or wiper duty cycle, the system directly measures the acceleration of objects traversing the road surface and uses this mechanical evidence to determine actual traction conditions, providing accurate real-time feedback regardless of environmental conditions.
3Reliability
If the AV controls maneuvering based on observed object acceleration, then the AV can determine minimum road surface traction for surfaces not yet traversed, but the system complexity increases due to fleet data communication requirements
Solution Approach 1:
The system uses a unified acceleration observation approach that serves multiple purposes: it determines traction for the individual vehicle's forward path, provides data for fleet-wide traction mapping, and enables both immediate vehicle control and long-term route planning. This multi-functional use of the same observation mechanism reduces overall system complexity despite the added reliability benefits.
Data Source
AI summary
Described herein are various technologies that pertain to controlling an AV based upon forward-looking observations of road surface behavior. With more specificity, technologies described herein pertain to controlling maneuvering of an AV in a region of a driving environment based upon an observed acceleration of an object in that region of the driving environment. A plurality of positions of an object in the driving environment are determined from output of sensors mounted on the AV. An acceleration of the object is computed based upon the positions. The AV is subsequently controlled based upon the computed acceleration.