Parking Occupancy UI Using Camera-Based Distance Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems rely on costly and error-prone sensors for navigation, leading to inaccurate detection and classification of moving and stationary objects, which affects the performance of autonomous or semi-autonomous driving.
Innovation Solution
A vision-based machine learning model that uses image sensors, such as cameras, to determine distances to objects relative to a robotic system, reducing the need for costly sensor hardware while enhancing accuracy through software complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If costly sensors are used for autonomous driving, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces complex mechanical sensor systems (LiDAR, radar, ultrasonic sensors) with a vision-based machine learning model that processes images from standard cameras. This substitution maintains measurement precision for object detection and distance determination while significantly reducing device complexity and hardware cost.
Solution Approach 2:
The patent creates a simulated 3D occupancy grid representation from 2D camera images through machine learning processing. This virtual copy of the physical environment enables accurate distance and object detection without requiring physical sensors that directly measure these parameters, thereby reducing hardware complexity while maintaining detection accuracy.
2Device complexity
If vision-based machine learning models are used, then device complexity is reduced, but measurement precision may worsen
Solution Approach 1:
The patent transforms 2D image data into a 3D occupancy grid representation through machine learning processing. This dimensional transformation enables the system to determine distances and spatial relationships with accuracy comparable to dedicated 3D sensors, while using only standard 2D camera hardware, thus resolving the precision- complexity tradeoff.
Solution Approach 2:
The patent changes the processing parameters from direct physical measurement (as done by LiDAR or radar) to statistical inference from image pixel data through trained neural networks. This parameter change allows the system to achieve accurate distance determination using software complexity instead of sophisticated hardware, reducing device complexity while maintaining measurement precision.
Data Source
AI summary
Systems and methods for enhanced user interface generation for parking based on occupancy machine learning model. An example method includes obtaining images from a multitude of image sensors positioned about a vehicle; computing a forward pass through an occupancy network to output, at least, information reflecting, for individual angular ranges about the vehicle, whether an object is within a threshold distance of the vehicle for an individual range along with an estimated distance to the object; and causing presentation, via a display of the vehicle, of a user interface depicting a graphical representation of the vehicle and the output information.