Vehicle MPC Parameter Learning for Adaptive Path Tracking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current model predictive controllers for autonomous vehicles rely on fixed operational, external, and control parameters, which do not account for dynamic changes in vehicle and environmental conditions, leading to suboptimal performance.

Innovation Solution

Implementing a machine learning-based method to dynamically update operational parameters, such as vehicle and external parameters, using simulations to determine optimal values for path tracking, corridor keeping, and collision avoidance maneuvers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If fixed operational parameters are used in model predictive controllers, then the controller structure is simple and easy to implement, but the controller cannot adapt to dynamic changes in vehicle and environmental conditions, leading to suboptimal performance

Engineering Contradiction:

Improveadaptability to dynamic changesVSAvoidcontroller complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic parameters that can change during operation based on vehicle conditions and environmental factors. The machine learning model continuously updates operational parameters (vehicle parameters, external parameters, and controls parameters) during simulation and real-time operation, transforming the static controller into a dynamic adaptive system that responds to changing conditions without requiring complex manual reconfiguration

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system employs self-learning through machine learning models that automatically adjust parameters based on simulated and real-world data. The controller performs self-updates by learning from operational data, eliminating the need for external manual parameter tuning and enabling autonomous adaptation to new conditions while maintaining manageable system complexity

Inventive Principle:

Principle #25Self-service

2Reliability

If learned operational parameters are used to adapt to dynamic changes, then the controller performance is improved, but the system complexity and computational requirements increase

Engineering Contradiction:

Improvecontroller performanceVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary learning during simulation phases before real-time deployment. The machine learning model is trained offline using simulated vehicle operations across various conditions, pre-learning optimal parameter values. This preliminary action transfers complex computational burden from real-time operation to offline simulation, maintaining high reliability during actual controller operation while managing real-time computational requirements

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces machine learning models as intermediary components between the fixed controller structure and dynamic operational conditions. These ML models act as adaptive layers that learn complex relationships without requiring fundamental changes to the underlying controller architecture, enabling performance improvement while containing system complexity through modular integration

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If fixed control parameters are used, then the computational load is low, but the path tracking and collision avoidance capabilities are suboptimal under varying conditions

Engineering Contradiction:

Improvepath tracking capabilityVSAvoidcomputational energy

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent applies different learning strategies to different parameter categories (vehicle parameters, external parameters, controls parameters) based on their specific characteristics and rates of change. This localized approach allows intensive learning for parameters that significantly impact path tracking performance while using simpler update mechanisms for less critical parameters, optimizing the balance between tracking capability and computational energy consumption

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS12559128B2Systems and methods for updating the parameters of a model predictive controller with learned operational and vehicle parameters generated using simulations and machine learning

Publication Date: 2026.02.24 TOYOTA JIDOSHA KK

AI summary

A computer implemented method for determining optimal values for operational parameters for a model predictive controller for controlling a vehicle can receive from a data store or a graphical user interface, ranges for one or more operational parameters. The computer implemented method can determine optimum values for vehicle parameters of the vehicle of one or more other parameters by simulating a vehicle operation across the ranges of the one or more operational parameters by solving a vehicle control problem and determining an output of the vehicle control problem based on a result for the simulated vehicle operation. A vehicle can include a processing component configured to adjust a control input for an actuator of the vehicle according to a control algorithm and based on the optimum values of a parameter as determined by the computer implemented method.