Side-View Camera Motion Detection for Safer Intersection Crossing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face challenges in accurately determining distances and speeds of moving and stationary objects, especially at intersections, due to limitations in sensor range and occlusions, which can lead to unsafe navigation decisions.
Innovation Solution
A camera vision method that uses side-view cameras to generate bounding boxes for objects, determine their direction and speed, and assess whether it is safe for the autonomous vehicle to move by analyzing a series of images over time, reducing reliance on expensive LiDAR sensors and improving detection accuracy at various road configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sensors (LiDAR) are used for object detection, then detection range is limited, but measurement precision deteriorates at distances beyond sensor range
Solution Approach 1:
The patent transitions from relying on active sensors like LiDAR to using passive camera-based vision systems that capture optical information from multiple dimensions and time points. By analyzing sequences of images from multiple camera angles, the system extends effective detection range beyond what single-point sensors can achieve, maintaining precision through multi-dimensional spatial and temporal analysis.
Solution Approach 2:
The system performs preliminary detection and tracking of objects using camera sequences before making navigation decisions. By continuously capturing and analyzing object positions, speeds, and trajectories in advance, the system builds predictive models that enable safe navigation decisions even when objects are at distances where traditional sensors lose precision.
2Measurement precision
If more sensors are added to extend detection range, then device complexity increases, but measurement precision may deteriorate due to sensor limitations
Solution Approach 1:
The patent employs camera systems that serve multiple functions: detecting object presence, determining position, calculating speed through frame-to-frame analysis, identifying direction of travel, and predicting future positions. This multi-functional approach eliminates the need for separate specialized sensors, reducing overall system complexity while maintaining or improving measurement precision across multiple parameters simultaneously.
Solution Approach 2:
The system uses image processing algorithms and computer vision techniques as intermediaries to extract precise measurement data from camera feeds. These computational intermediaries transform raw visual information into accurate object parameters (position, speed, direction), achieving precision comparable to or exceeding direct sensor measurements without adding physical sensor complexity.
3Reliability
If sensor coverage is increased to detect all objects, then device complexity increases, but reliability deteriorates due to occlusions and sensor limitations
Solution Approach 1:
The patent divides the detection task into segments handled by multiple cameras positioned at different locations. Each camera captures a specific field of view, and the system integrates data from these segmented views to build a complete picture of the environment. This segmentation approach reliably detects objects that might be occluded from any single viewpoint while keeping each individual camera simple and reliable.
Solution Approach 2:
The system continuously monitors object positions and movements, using feedback from sequential image analysis to update predictions of object trajectories. This feedback loop allows the system to adapt to changing conditions, detect objects that enter or leave the scene, and maintain reliable detection even when objects temporarily occlude each other, as the continuous stream of data provides redundant information.
Data Source
AI summary
A method includes receiving a series of road images from a side-view camera sensor of the autonomous driving vehicle. For each object from objects captured in the series of road images, a series of bounding boxes in the series of road images is generated, and a direction of travel or stationarity of the object is determined. The methods and apparatus also include determining a speed of each object for which the direction of travel has been determined and determining, based on the directions of travel, speeds, or stationarity of the objects, whether the autonomous driving vehicle can safely move in a predetermined direction. Furthermore, one or more control signals is sent to the autonomous driving vehicle to cause the autonomous driving vehicle to move or to remain stationary based on determining whether the autonomous driving vehicle can safely move in the predetermined direction.