Adaptive Steering Feedback for Autonomous Vehicle Oscillation Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicle steering control systems experience high amplitude oscillations due to lack of adaptation mechanisms for actuator and environmental uncertainties, which are not addressed by existing gain scheduled feedback control strategies.
Innovation Solution
A model-reference adaptive steering control strategy that includes an adaptive feedback methodology to automatically adjust steering feedback nonlinearity, using equations to determine a feedback gain correction vector and torque overlay, enhancing steering control performance and reducing calibration effort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If gain scheduled feedback control is used, then steering control is implemented, but high amplitude oscillations occur due to lack of adaptation mechanism
Solution Approach 1:
The patent implements a model-reference adaptive control system that uses feedback to continuously adjust steering control parameters. The feedback mechanism compares actual steering behavior with desired reference models and automatically corrects deviations, enabling the system to adapt to uncertainties in real-time while maintaining stability and reducing oscillations.
Solution Approach 2:
The system dynamically adjusts control parameters during operation rather than using fixed gain schedules. The adaptive controller modifies its behavior based on real-time conditions, allowing the steering system to respond flexibly to changing actuator characteristics and environmental factors, thereby eliminating high amplitude oscillations while maintaining reliability.
2Reliability
If gain scheduled control strategy is used, then steering control is achieved, but calibration effort increases due to need for multiple scenario calibrations
Solution Approach 1:
The adaptive control system performs self-calibration during normal operation by automatically learning and adapting to actuator characteristics and environmental conditions. Rather than requiring manual calibration across multiple scenarios, the system serves itself by continuously adjusting its parameters based on observed behavior, significantly reducing calibration effort while maintaining control performance.
Solution Approach 2:
The system automatically changes control parameters during operation based on real-time conditions. The adaptive algorithm modifies gain values and other parameters dynamically, eliminating the need for pre-calibration across different scenarios. This parameter adaptation occurs seamlessly during normal vehicle operation, reducing manufacturing and calibration complexity.
3Reliability
If traditional feedback control is used, then steering control is implemented, but steering feedback proliferation occurs
Solution Approach 1:
The model-reference adaptive controller serves multiple functions within a unified framework. It simultaneously handles steering control, adaptation to uncertainties, and parameter optimization, eliminating the need for separate feedback mechanisms for each function. This multi-functionality reduces feedback proliferation while maintaining stability and improving overall system efficiency.
Data Source
AI summary
An adaptive method for controlling the steering of an autonomous vehicle includes receiving a steering input. The steering input is indicative of a desired road wheel angle. The method also includes determining a commanded torque to achieve the desired road wheel angle. The method also includes measuring the road wheel angle of the vehicle (i.e., the measured road wheel angle). The method also includes executing a feedback control to determine a torque overlay. The feedback control includes determining, in real time, a current steering error vector of the vehicle. The feedback control includes determining the difference between the current steering error vector and a predetermined-steering error dynamics vector. The feedback control also includes determining a feedback gain correction vector. The feedback gain correction vector is a function of the difference between the current steering error vector and the predetermined-steering error dynamics vector.