V2V Blind-Spot Warning Using CNN Video Detection in Low Visibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional blind-spot warning systems face challenges in providing accurate warnings due to deteriorated position accuracy in adverse driving conditions, camera contamination, and low illumination, especially at night, using GPS and V2V communication.
Innovation Solution
A method employing convolutional neural networks and V2V communication to analyze front and rear videos from cameras, generating feature maps and feature vectors to detect vehicles in blind spots, and providing warnings based on longitudinal and lateral distances, even in extreme conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS-based V2V communication is used for blind-spot warning, then position information can be shared between vehicles, but position accuracy deteriorates in adverse driving conditions such as between tall buildings, cloudy weather, tunnels, and nighttime
Solution Approach 1:
The patent introduces an intermediary object detection system using cameras and neural networks to mediate between GPS-based V2V communication and blind-spot warning. When GPS position accuracy deteriorates in adverse conditions, the system uses camera-based object detection as an intermediary to identify vehicles in blind spots, thereby maintaining warning reliability without direct dependence on precise GPS positioning.
Solution Approach 2:
The system dynamically changes operational parameters by switching between GPS-based warning mode and camera-based object detection mode depending on environmental conditions. When GPS accuracy degrades (parameter change in positioning reliability), the system activates alternative detection parameters through neural network analysis of camera images, adapting to maintain overall system performance.
2Reliability
If rear camera is used for object detection, then blind-spot monitoring can be provided, but detection performance is low when the lens is contaminated with water-drops or during nighttime driving in low illumination
Solution Approach 1:
The patent applies partial action by using only the front camera when rear camera detection is compromised. Instead of requiring both cameras to function perfectly, the system can operate with partial detection capability using front camera images processed through neural networks to identify vehicles in blind spots, maintaining adequate detection reliability under adverse conditions.
Solution Approach 2:
The system creates a virtual copy of rear-view detection capability by using the front camera to capture images, which are then processed through neural network algorithms to infer the presence of vehicles in blind spots. This copying approach allows the system to maintain detection functionality even when the actual rear camera is contaminated or ineffective.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for giving a warning on a blind spot of a vehicle based on V2V communication is provided. The method includes steps of: (a) if a rear video of a first vehicle is acquired from a rear camera, a first blind-spot warning device transmitting the rear video to a blind-spot monitor, to determine whether nearby vehicles are in the rear video using a CNN, and output first blind-spot monitoring information of determining whether the nearby vehicles are in a blind spot; and (b) if second blind-spot monitoring information of determining whether a second vehicle is in the blind spot, is acquired from a second blind-spot warning device of the second vehicle, over the V2V communication, the first blind-spot warning device warning that one of the second vehicle and the nearby vehicles is in the blind spot by referring to the first and the second blind-spot monitoring information.