Self-Driving Car Steering Profiles Within Physical Steering Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving cars face challenges in planning trajectories around obstacles and controlling steering to safely transition between predetermined states, with existing systems often requiring steering wheel positions that are either dangerous or outside the physically possible range.
Innovation Solution
A computer-implemented method using a processor to determine a target steering profile for a Self-Driving Car (SDC) by acquiring SDC state data, building a polynomial curve representing candidate steering positions, and simulating a trajectory to select a profile within the permissible range, while training a Machine Learning Algorithm (MLA) to predict in-use steering wheel positions for safe operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a steering profile is generated to transition the SDC from one predetermined state to another, then the trajectory planning is achieved, but the steering wheel positions may fall outside the permissible range or be physically impossible
Solution Approach 1:
The system dynamically adjusts the steering profile by iteratively modifying steering wheel positions based on simulation feedback. The polynomial curve parameters are adjusted in real-time to ensure steering positions remain within permissible ranges while achieving the desired trajectory transition between states.
Solution Approach 2:
The system employs feedback loops where simulated trajectories are evaluated against permissible steering ranges, and the polynomial curve parameters are adjusted based on this feedback. The simulation results feed back into the parameter adjustment process, iteratively refining the steering profile until it satisfies both trajectory and steering constraints.
2Reliability
If the SDC follows a planned trajectory to avoid obstacles, then collision risk is reduced, but the steering positions required may be dangerous or outside physical limits
Solution Approach 1:
The system performs preliminary simulation of the steering profile before actual execution. By simulating the trajectory and evaluating steering positions in advance, the system identifies and corrects dangerous or impossible steering positions before they are applied to the vehicle, ensuring safety and feasibility.
Solution Approach 2:
The system incorporates safety margins and constraints in the polynomial curve parameter selection to prevent dangerous steering positions. By pre-defining permissible steering ranges and using these as boundaries in the optimization process, the system cushions against generating unsafe steering commands.
3Productivity
If polynomial curve parameters are selected to define steering profiles, then trajectory transitions are enabled, but the parameters may result in steering positions outside the permissible range
Solution Approach 1:
The system changes the parameters of the polynomial curve iteratively to achieve the desired balance between trajectory efficiency and steering position accuracy. By adjusting polynomial coefficients and exponents based on simulation feedback, the system optimizes both the speed of state transitions and the accuracy of steering positions within permissible ranges.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and processor for controlling steering of a Self-Driving Car (SDC) are disclosed. The method includes: acquiring SDC state data associated with vertices in a graph-structure, using the SDC state data for building a polynomial curve, using the polynomial curve for simulating a candidate trajectory of the SDC as if the SDC is steered in accordance therewith and such that a closest allowed steering wheel position to a non-allowed steering wheel position is used instead of the non-allowed steering wheel position, determining an offset between the simulated state and a target state of the SDC, selecting candidate steering profile as a target steering profile of the SDC based on the offset, and using the target steering profile of the SDC for controlling steering of the SDC. A method of training a Machine Learning Algorithm for predicting steering wheel positions to control steering of the SDC is also disclosed.