Lateral Positioning Control for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in determining safe lateral positioning within a lane, as following the geometric center may not be sufficient, especially considering road curvature, static and dynamic obstacles, and varying object speeds, which existing systems fail to adequately address.
Innovation Solution
A computing device identifies and characterizes objects along the vehicle's trajectory, estimates time intervals of lateral adjacency, determines optimal lateral distances based on object characteristics and vehicle speed, and adjusts the trajectory to maintain safe distances, using heuristics and weights to mimic human-like behavior, while considering road conditions and lane types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle follows the geometric center of the lane, then the lateral positioning is simple to implement, but the safety distance from obstacles is not adequately maintained
Solution Approach 1:
The system pre-calculates safe lateral distance thresholds for different object types (vehicles, cyclists, pedestrians, animals) and stores them as reference data. During operation, the controller directly retrieves and applies these pre-determined safety distances without complex real-time calculations, ensuring reliable safety margins while maintaining system simplicity
Solution Approach 2:
The trajectory is dynamically adjusted based on real-time detection of objects and their characteristics. The controller modifies the geometric center trajectory by adding lateral offset distances that are determined by object type, creating a dynamic safe trajectory that adapts to varying environmental conditions while building upon the simple geometric center baseline
2Productivity
If the vehicle maintains a fixed lateral distance from objects, then the control logic is simple, but the navigation efficiency deteriorates due to unnecessary trajectory adjustments
Solution Approach 1:
Different lateral distance thresholds are applied locally based on object characteristics. Instead of using a uniform safety distance, the system assigns specific distance values to different object types (e.g., larger distances for animals and cyclists, smaller distances for fixed infrastructure), allowing efficient navigation while maintaining appropriate safety margins for each local situation
Solution Approach 2:
The system changes the safety distance parameter dynamically based on object classification. When an object is detected, the controller retrieves the corresponding distance threshold from stored reference data and applies it to adjust the trajectory, optimizing navigation efficiency by avoiding excessive conservatism while ensuring safety through parameter adaptation
3Adaptability or versatility
If the system considers multiple object characteristics and environmental factors, then the safety and adaptability are improved, but the computational complexity increases
Solution Approach 1:
The control system is segmented into distinct functional modules: object detection module, object classification module, distance determination module (using lookup tables), and trajectory adjustment module. Each module handles a specific aspect of the control task, making the overall complex system manageable and maintainable while achieving high adaptability to environmental factors
Solution Approach 2:
Reference data tables serve as intermediaries between object detection and trajectory control. The tables store pre-calculated safe distance thresholds for various object types and conditions, acting as a mediator that translates complex safety requirements into simple lookup operations, reducing computational complexity while maintaining adaptability
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems for controlling a vehicle are described. A computing device may be configured to identify an object on or proximate to a road on which a vehicle is travelling. The computing device may be configured to estimate, based on characteristics of the vehicle and respective characteristics of the object, one or more positions at which the vehicle will be adjacent to the object, wherein the characteristics of the vehicle include a speed and a direction of motion of the vehicle, and wherein the respective characteristics of the object include a respective speed and a respective direction of motion of the object. Based on the respective characteristics of the object, the computing device may be configured to determine a distance for the vehicle to maintain between the vehicle and the object at the one or more positions of the vehicle; and provide instructions to control the vehicle based on the distance.