Autonomous Vehicle Steering Rate Optimization via Cost Function Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion planning and control methods for autonomous vehicles do not accurately consider the differences in vehicle types, leading to potential dangers and discomfort due to high steering rates and overshooting during navigation.
Innovation Solution
A computer-implemented method that determines a target steering rate by calculating individual costs for various steering rate candidates using cost functions based on differences in target and current steering positions, vehicle speed, and past steering data, selecting the candidate with the lowest total cost to generate a steering control command.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a high steering rate is used to achieve steer control accuracy, then the steering response speed is improved, but the safety of the vehicle steering system and passengers deteriorates
Solution Approach 1:
The patent applies dynamics by making the steering rate adjustable rather than fixed. The system dynamically selects from multiple steering rate candidates (e.g., 45 deg/s, 90 deg/s, 135 deg/s, 180 deg/s) based on real-time conditions. This allows the steering rate to adapt to different situations, achieving high accuracy when needed while maintaining safety by reducing the rate when conditions warrant caution.
Solution Approach 2:
The patent changes the parameter of steering rate from a single fixed value to multiple selectable values. By evaluating cost functions for different steering rate candidates and selecting the optimal one, the system transforms the steering rate from a static parameter to a dynamic one that can be optimized for both accuracy and safety based on current vehicle state and environmental conditions.
2Speed
If a high steering rate is used to achieve steer control accuracy, then the steering response speed is improved, but the risk of overshooting deteriorates
Solution Approach 1:
The patent implements feedback by continuously evaluating the difference between target and actual steering angles, as well as considering vehicle speed and acceleration. This feedback loop allows the system to adjust the steering rate dynamically, reducing it when approaching the target position or when vehicle dynamics suggest a high rate would cause overshooting, thereby maintaining both fast response and accuracy.
Solution Approach 2:
The system dynamically adjusts the steering rate based on real-time vehicle state and positioning information. By selecting from multiple steering rate candidates based on current conditions, the system achieves fast response when appropriate while preventing overshooting by reducing the rate when conditions warrant caution.
3Device complexity
If same motion planning and control is applied to all types of vehicles, then the system complexity is reduced, but the accuracy and smoothness of navigation deteriorates
Solution Approach 1:
The patent applies local quality by tailoring the motion planning and control parameters to specific vehicle types. Instead of using a one-size-fits-all approach, the system considers vehicle-specific characteristics (such as steering geometry, mass, and dynamics) when selecting steering rate candidates and evaluating cost functions. This allows each vehicle type to receive optimized control parameters suited to its characteristics, improving navigation accuracy and smoothness.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one embodiment, a number of steering rate candidates are determined for a steering control command of operating an autonomous vehicle. For each of the steering rate candidates, a number of individual costs are calculated for the steering rate candidate by applying a plurality of cost functions, each cost function corresponding to one of a plurality of cost categories. A total cost for the steering rate candidate is determined based on the individual costs produced by the cost functions. One of the steering rate candidates having a lowest total cost is selected as a target steering rate. A steering control command is generated based on the selected steering rate candidate to control a steering wheel of the autonomous vehicle.