Camera-Based Vehicle Steering via Travel Potential Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control systems face high processing loads due to integrating outputs from multiple external sensors and map information, which hinders efficient operation in sustainable transportation systems.
Innovation Solution
A vehicle control device that generates a travel potential field and target trajectory solely based on images from a forward-facing camera, using reference object information to set low-potential and high-potential points, reducing the need for external sensors and map data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple external sensors (camera, radar, LiDAR) and map information are integrated to recognize travel path and surrounding objects, then the accuracy and completeness of environmental recognition is improved, but the processing load on the in-vehicle computer increases significantly
Solution Approach 1:
The patent extracts and utilizes only the camera sensor data for generating the travel potential field, eliminating the need to process and integrate data from radar, LiDAR, and map information. This selective extraction reduces the processing load while maintaining sufficient accuracy for safe vehicle operation through the potential field method.
Solution Approach 2:
The camera image is transformed into a universal representation (travel potential field) that encodes both the travel path and surrounding obstacle information. This single image-based potential field serves multiple functions: path planning, obstacle detection, and navigation control, replacing the need for separate processing of multiple sensor types.
2Adaptability or versatility
If the in-vehicle computer integrates outputs from multiple external sensors and map information to generate the potential function, then the comprehensiveness of travel planning is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent removes radar, LiDAR, and map data from the computational process, relying solely on camera images to construct the travel potential field. This extraction of unnecessary data sources simplifies the computational complexity while maintaining adaptable travel planning through the image-based potential field representation.
Solution Approach 2:
The camera image serves as a simplified copy or representation of the real-world environment, from which the travel potential field is derived. This image-based copy contains sufficient information for comprehensive travel planning without requiring the complex processing and integration of multiple original sensor data streams.
3Productivity
If a travel potential field is generated using only camera images and reference object information, then the processing load is reduced, but the amount of available information for navigation decisions decreases
Solution Approach 1:
The camera image creates a simplified copy of the environment that is sufficient for navigation decisions. From this copy, the travel potential field is extracted, which contains all necessary navigation information (path and obstacles) without requiring the additional information processing that would be needed from multiple sensor types.
Solution Approach 2:
The patent transforms the camera image into a travel potential field by changing the representation parameters from raw pixel data to a potential field distribution. This parameter transformation consolidates navigation information into a compact representation that is sufficient for decision-making while requiring minimal processing resources.
Data Source
AI summary
A vehicle control device generates a travel potential field based on an input image, generates a target trajectory based on a gradient of the travel potential, and operates an electric power steering device based on the target trajectory. The vehicle control device executes the processing of: setting a low-potential point at a position within a central region of the input image, the position being determined based on reference object information; setting a first high-potential point in a region of the input image where an obstacle appears; setting a value of the travel potential at the low-potential point to a first set value; setting a value of the travel potential at the first high-potential point to a second set value; and generating the travel potential field by interpolating a value of the travel potential, in a region between the low-potential point and the first high-potential point in the input image.