Vehicle Trajectory Feed-Forward Control Without Inverse-Dynamics Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for controlling vehicle longitudinal and lateral guidance, especially in automated driving systems, face challenges with non-invertible vehicle dynamics, leading to complex filter designs and delayed responses, which hinder efficient feed-forward control implementation.
Innovation Solution
An electronic device and method that utilize a feed-forward controller to calculate actuation inputs based on setpoint trajectory variables and a vehicle-specific dynamic model, eliminating the need for filters and reducing complexity, thereby enabling efficient control even with non-invertible vehicle dynamics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If model-based feed-forward control is used to control vehicle longitudinal and lateral guidance, then the vehicle can follow the planned trajectory with consideration of system dynamics, but the controller becomes complex and difficult to implement when dealing with non-invertible vehicle dynamics
Solution Approach 1:
The controller is segmented into three distinct modules: a trajectory generator that creates reference trajectories, a feed-forward controller that calculates control inputs from trajectory derivatives, and a feedback controller that corrects deviations. This segmentation allows each module to have a specific, simplified function, reducing overall controller complexity while maintaining trajectory following accuracy
Solution Approach 2:
The feed-forward controller calculates control inputs in advance based on the desired trajectory and vehicle dynamics model, before actual trajectory execution. By pre-calculating the required control actions using the relationship between trajectory derivatives and control inputs, the system prepares control commands that account for vehicle dynamics without requiring complex real-time inversion during execution
2Stability of the object's composition
If filters are applied to handle non-invertible vehicle dynamics, then the controller can be stabilized, but delay times increase and response speed decreases
Solution Approach 1:
The system pre-calculates control inputs using the feed-forward controller based on the desired trajectory and vehicle dynamics, obtaining control values in advance without requiring filtering operations. This preliminary calculation approach provides stable control signals before they are applied, eliminating the need for stabilizing filters that would introduce delays
Solution Approach 2:
The trajectory generator acts as an intermediary that transforms the desired trajectory into a form suitable for the feed-forward controller. By generating reference trajectories with appropriate smoothness and continuity properties, this intermediary component ensures that subsequent control calculations remain stable without requiring additional filtering
3Stability of the object's composition
If complex filter designs are used to handle non-invertible dynamics, then controller stability can be achieved, but the implementation becomes more difficult and computationally intensive
Solution Approach 1:
By dividing the control system into separate trajectory generation and control calculation modules, each with well-defined functions, the implementation becomes more straightforward. The trajectory generator handles trajectory planning with guaranteed smoothness properties, while the feed-forward controller handles control input calculation, eliminating the need for complex integrated filter designs
Solution Approach 2:
The system performs preliminary trajectory generation and control calculation in a structured sequence, where the trajectory generator first establishes a smooth reference trajectory, and then the feed-forward controller calculates control inputs. This preliminary structured approach ensures stability without requiring complex filter designs, making implementation easier
Data Source
AI summary
Methods, systems, and apparatuses are provided for controlling longitudinal and/or lateral guidance of a vehicle. A trajectory planner is configured to ascertain setpoint trajectory variables for a trajectory of the vehicle. The setpoint trajectory variables include a first trajectory state and a second trajectory state. The second trajectory state corresponds to a change in the first trajectory state over time. A feed-forward controller is configured to calculate a feed-forward control variable from the setpoint trajectory variables and from a model of dynamic behavior of the vehicle. The feed-forward control variable and the first trajectory state are variables of the same type. The device is configured to use the feed-forward control variable as a basis for determining a manipulated variable for a controlled system within the scope of the longitudinal and/or lateral guidance. The manipulated variable and the feed-forward control variable are variables of the same type.