V2V Object Detection Fusion for Autonomous Vehicle Safety
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicles face challenges in sharing space detection results effectively due to each vehicle performing its own spatial detection independently, which hinders the improvement of safety through vehicle-to-vehicle communication.
Innovation Solution
A learning method and device that integrate object detection information from other autonomous vehicles via V2V communication with the object detection information generated by the present autonomous vehicle, using a merging network to combine and enhance the detection results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If each autonomous vehicle performs its own spatial detection independently, then the vehicle can maintain its own detection accuracy, but the safety improvement through V2V communication cannot be achieved
Solution Approach 1:
The patent merges object detection results from multiple autonomous vehicles through V2V communication into a unified detection result. The server collects detection results from multiple vehicles and combines them to generate comprehensive spatial detection information, thereby improving overall safety while eliminating information loss through independent detection.
2Measurement precision
If object detection information from multiple vehicles is integrated, then the safety and detection accuracy improve, but the complexity of the system increases
Solution Approach 1:
The patent introduces a server as an intermediary to handle the complex task of integrating detection information from multiple vehicles. The server receives detection results from various vehicles, performs the integration and matching operations, and returns the unified detection result to the requesting vehicle, thereby managing system complexity centrally rather than distributing it across all vehicles.
Solution Approach 2:
The system segments the detection integration process into distinct modules: detection result generation by individual vehicles, collection by the server, matching based on spatial-temporal relationships, and unified result generation. This segmentation allows each component to focus on a specific task, managing overall system complexity while achieving high detection accuracy.
3Reliability
If V2V communication is used to share detection results, then the probability of accidents is reduced, but the difficulty of sharing spatial detection results increases due to different detection perspectives
Solution Approach 1:
The patent performs preliminary actions by having each vehicle generate detection results in advance and transmit them to the server before integration is needed. The server pre-processes these results by organizing them according to spatial-temporal relationships, so when a vehicle needs integrated detection information, the matching and integration have already been prepared, reducing the difficulty of real-time integration despite different detection perspectives.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A learning method for generating integrated object detection information by integrating first object detection information and second object detection information is provided. And the method includes steps of: (a) a learning device instructing a concatenating network to generate one or more pair feature vectors; (b) the learning device instructing a determining network to apply FC operations to the pair feature vectors, to thereby generate (i) determination vectors and (ii) box regression vectors; (c) the learning device instructing a loss unit to generate an integrated loss by referring to the determination vectors, the box regression vectors and their corresponding GTs, and performing backpropagation processes by using the integrated loss, to thereby learn at least part of parameters included in the DNN.