Vehicle Motion Control via Invariant Set Trajectory Coordination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advanced driver assistance (ADA) and autonomous driving (AD) systems face challenges in coordinating vehicle control components to accurately follow a desired trajectory, due to differences in motion models used for trajectory generation and control, which can lead to inaccuracies in executing complex vehicle motions like lane keeping and collision avoidance.
Innovation Solution
The system employs a method that selects a first motion model for trajectory generation and a higher-order second motion model for control, using a control invariant set to ensure mutual dependency and precise vehicle control, with constraints imposed on the state of the vehicle to maintain performance metrics such as safety margins and trajectory adherence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a simplified motion model is used for trajectory generation, then computational speed is improved, but trajectory execution accuracy deteriorates
Solution Approach 1:
The control system is divided into two segments: a supervisory controller that generates trajectories using a simplified motion model for computational efficiency, and a vehicle controller that executes trajectories using a more accurate higher-order motion model for precision. This segmentation allows each component to use the appropriate level of model complexity for its function.
Solution Approach 2:
A control invariant set acts as an intermediary between the simplified trajectory generation and the accurate trajectory execution. It ensures that trajectories generated by the simplified model can be successfully executed by the more accurate model, bridging the gap between computational speed and execution accuracy requirements.
2Manufacturing precision
If a higher-order motion model is used for control, then trajectory execution accuracy is improved, but computational complexity increases
Solution Approach 1:
The higher-order motion model is applied locally only during the trajectory execution phase where precision is critical, rather than throughout the entire control process. This allows high accuracy where needed while maintaining overall system efficiency.
Solution Approach 2:
The supervisory controller performs preliminary trajectory generation using the simplified model, preparing trajectories that are guaranteed to be executable by the more accurate model through the control invariant set. This preliminary action reduces the computational burden on the execution phase.
3Productivity
If different motion models are used for trajectory generation and control, then computational efficiency is improved, but coordination difficulty increases
Solution Approach 1:
The control invariant set serves as a formal intermediary that mathematically guarantees compatibility between trajectories generated by the simplified model and executed by the accurate model. This formal framework systematically manages the coordination between different models.
Solution Approach 2:
The system uses feedback from the vehicle's actual state (measured by sensors) to the vehicle controller, which adjusts control commands to ensure the vehicle follows the desired trajectory despite model differences. This feedback loop compensates for discrepancies between the two motion models.
Data Source
AI summary
A method selects from a memory a first model of motion of vehicle, a second model of the motion of the vehicle, a first constraint on the first model for moving along a desired trajectory of the vehicle, and a control invariant set joining states of the first model with states of the second model. For each combination of the states within the control invariant subset there is at least one control action to the second model that maintains the state of the second model within the control invariant set for every modification of the state of the first model satisfying the first constraint. A portion of the desired trajectory satisfying the first constraint is determined using the first model while a sequence of commands for moving the vehicle along the portion of the desired trajectory is determined using the second model. The sequence of commands is determined to maintain the sequence of the states of the second model and a sequence of the states of the first model determined by the portion of the desired trajectory within the control invariant subset. The vehicle is controlled using at least one command from the sequence of commands.