Related Experiment Video
Updated: Apr 18, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.2K
Edge-assisted adaptive offloading algorithm for 3D object detection tasks
Kangli Zhao1,2, Zhongrui Gou2, Huaqing Liu1
1School of Computer Science and Technology, Aba Teachers University, Aba, Sichuan, China.
Plos One
|April 16, 2026
Summary
This study introduces an edge computing framework to reduce delay in multimodal 3D object detection for autonomous systems. The approach optimizes performance by balancing computational load, achieving a better delay-accuracy trade-off.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Edge Computing
Background:
- Multimodal 3D object detection is essential for autonomous systems.
- High computational demands lead to significant latency, hindering real-time performance.
Purpose of the Study:
- To develop an edge computing-assisted framework to reduce delay in multimodal 3D object detection.
- To optimize the balance between computational load on terminal devices and edge servers.
Main Methods:
- Implemented a framework that offloads computation between devices and edge servers.
- Introduced dynamic threshold tuning and resolution-adaptive offloading algorithms.
- Evaluated performance based on delay, accuracy, and adaptability.
Main Results:
- Significantly reduced detection delay by minimizing offloading frequency.
- Maintained high detection accuracy, achieving a superior delay-accuracy trade-off.
- Demonstrated robust adaptability across different models and bandwidth conditions.
Conclusions:
- The proposed edge computing framework effectively reduces latency in multimodal 3D object detection.
- The dynamic and adaptive algorithms ensure efficient performance in diverse operational environments.
- This approach offers a practical solution for enhancing autonomous system capabilities.