Autonomous Truck Control Modeling Using Sparse Data Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional modeling techniques for autonomous vehicles, such as semi-trailer trucks, are overly complex and require manual adjustments to match vehicle operations, leading to inefficiencies and time-consuming tuning processes due to mismatches between derived equations of motion and actual vehicle behavior.
Innovation Solution
A data-driven control system using a system identification process and sparse regression techniques to derive a control model from vehicle data, allowing for the development of a robust feedback controller that minimizes errors caused by disturbances, such as wind or potholes, by constructing matrices from sampled data and solving optimization problems to determine non-zero terms that dictate the control model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional modeling techniques are used to derive equations of motion for autonomous vehicles, then the control model can be obtained through theoretical derivation, but the model becomes overly complex and requires manual adjustments to match actual vehicle operations
Solution Approach 1:
The patent replaces conventional theoretical modeling approaches with data-driven system identification techniques. Instead of deriving equations of motion through mechanical analysis and manual parameter tuning, the system uses measured vehicle data to automatically identify control model parameters through optimization algorithms, thereby reducing model complexity while maintaining accuracy
Solution Approach 2:
The control model automatically adapts to the specific vehicle through self-identification using onboard sensor data. The system performs autonomous parameter identification without requiring external manual adjustments, allowing the model to self-tune to the actual vehicle characteristics through iterative optimization
2Reliability
If conventional modeling techniques are used, then theoretical control models can be derived, but manual adjustments and tuning processes become time-consuming
Solution Approach 1:
The system performs control model identification during normal vehicle operation before actual autonomous driving deployment. By collecting and processing operational data in advance, the system prepares the optimized control parameters beforehand, eliminating the need for time-consuming manual tuning during deployment phases
Solution Approach 2:
The system continuously compares predicted vehicle behavior with actual measured behavior and uses this feedback to iteratively refine control model parameters. This closed-loop identification process automatically converges to optimal parameters without manual intervention, significantly reducing tuning time
3Device complexity
If data-driven approaches are used with sparse regression techniques, then the control model complexity is reduced and tuning requirements are minimized, but robust feedback controllers must be designed to handle disturbances and uncertainties
Solution Approach 1:
The system transforms the control approach by changing from fixed theoretical parameters to data-identified parameters that are optimized for the specific vehicle. Sparse regression techniques identify only the most significant parameters, reducing model complexity while maintaining or improving robustness through empirical validation
Solution Approach 2:
The patent replaces traditional robust control design methods with data-driven identification followed by modern controller synthesis. By using identified parameters from sparse regression, the system achieves robustness through accurate vehicle-specific modeling rather than conservative theoretical designs
Data Source
AI summary
Techniques are described to determine parameters and/or values for a control model that can be used to operate an autonomous vehicle, such as an autonomous semi-trailer truck. For example, a method of obtaining a data-driven model for autonomous driving may include obtaining data associated with a first set of variables that characterize movements of an autonomous vehicle over time and commands provided to the autonomous vehicle over time, determining, using at least the first set of data, non-zero values and an associated second set of variables that describe a control model used to perform an autonomous driving operation of the autonomous vehicle, and calculating values for a feedback controller that describes a transfer function used to perform the autonomous driving operation of the autonomous vehicle driven on a road.