Single-Camera Path Planning With Semantic Cost Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle navigation systems that rely on stereo camera or multi-camera systems are bulky, power-intensive, and require both cameras to function, leading to slow scene interpretation and potential system failure if one camera malfunctions, limiting speed and efficiency.
Innovation Solution
Utilizing a single image sensor for path planning with semantic segmentation and machine learning to identify hazards, projecting images onto a 2D map, and generating a cost map for efficient path determination, which can also serve as a backup or supplement to stereo systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple cameras are used for depth perception, then navigation reliability is improved, but device complexity and weight increase
Solution Approach 1:
The patent segments the navigation function into two parts: a primary multi-camera system for normal operation and a secondary single-camera system for fail-safe operation. This segmentation allows the system to maintain high reliability through redundancy while managing device complexity by having simpler backup systems.
Solution Approach 2:
The patent changes the operational parameters of the image sensor by adjusting exposure time, gain, and resolution settings based on vehicle speed and lighting conditions. This allows a single sensor to adapt to different operational requirements, reducing the need for multiple specialized sensors.
2Reliability
If multiple cameras are used for depth perception, then navigation reliability is improved, but power consumption increases
Solution Approach 1:
The patent implements dynamic switching between single-camera and multi-camera modes based on operational conditions. The system activates additional cameras only when needed for fail-safe operation or complex navigation scenarios, reducing overall power consumption while maintaining reliability when required.
Solution Approach 2:
The patent dynamically adjusts image sensor parameters such as exposure time, gain, and resolution based on lighting conditions and vehicle speed. This allows the system to maintain navigation reliability while minimizing power consumption by using lower resolution and shorter exposure times in favorable conditions.
3Measurement precision
If image sensor parameters are increased for low-light conditions, then scene detection accuracy is improved, but power consumption increases
Solution Approach 1:
The patent dynamically adjusts image sensor parameters including exposure time, gain, and resolution based on ambient lighting conditions. In low-light conditions, the system increases exposure time and gain to maintain detection accuracy while avoiding excessive power consumption by only adjusting parameters when necessary.
Solution Approach 2:
The patent implements real-time monitoring of lighting conditions and dynamically switches between different sensor configurations. The system activates high-power modes only when scene detection accuracy is compromised by lighting conditions, thereby balancing measurement precision with power consumption.
4Device complexity
If a single image sensor is used, then device complexity is reduced, but reliability decreases due to lack of redundancy
Solution Approach 1:
The patent makes the single image sensor universal by implementing multiple operational modes and functions within a single device. The sensor can operate in high-resolution mode, low-power mode, and fail-safe mode, effectively replacing multiple specialized sensors while maintaining system reliability through software-based redundancy and adaptive processing.
Data Source
AI summary
Systems and methods for autonomous vehicle path planning are described herein. An example vehicle includes an image sensor to obtain an image of a scene of an area surrounding the vehicle. The vehicle also includes navigation system circuitry to: analyze the image and generate a semantically segmented image that identifies one or more types of features in the image; project the semantically segmented image to a two-dimensional (2D) map projection; convert the 2D map projection into a cost map; and determine a path for the vehicle based on the cost map.