Collaborative Automated Driving: A Machine Learning-based Method to Enhance the Accuracy of Shared Information

Rawashdeh, Zaydoun Yahya; Wang, Zheng

doi:10.1109/itsc.2018.8569832

Cited by 36 publications

(22 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite many efforts and successes in developing dynamic map fusion techniques [4]- [15], a number of technical challenges still need to be properly handled, including 1) Reduction of sensing uncertainties. The sensing data at each individual vehicle may be missing, noisy, or mistaken due to sensor limitation and environmental complexity.…”

mentioning

confidence: 99%

“…Current dynamic map fusion techniques can be classified as data level [4], [9], feature level [5]- [8], and object level [9]- [15]. Those data-level and feature-level methods likely incur high communication overheads and most object-level methods do not take uncertainties into account.…”

mentioning

confidence: 99%

“…Those data-level and feature-level methods likely incur high communication overheads and most object-level methods do not take uncertainties into account. On the other hand, various deep and federated learning schemes have been used to improve the quality of feature models [8], [14], [15], but few of them use point-cloud data. Besides, as most cloud-based model training schemes assume that datasets have been properly labeled [4]- [9], [11], [14], [15], knowledge distillation (KD) methods [16]- [18] need further investigation for label generation among networked vehicles.…”

mentioning

confidence: 99%

“…The labels of real-time sensory data are provided by road-side units or generated from the three-stage fusion algorithm at the edge server. feature level [5]- [8], and object level [9]- [15]. Data level methods collect raw data (such as point clouds) from different vehicles and achieve global dynamic map fusion by using deep networks [4], [9] at the cost of high communication overhead and extra data synchronization.…”

mentioning

confidence: 99%

“…Feature level methods deploy identical feature extraction models (such as voxelization [5]- [7] and segmentation [8]) at different vehicles, and achieve global dynamic map fusion via feature collection and feature processing based on deep networks. Object level methods usually share the object lists among vehicles and achieve global dynamic map fusion through object association and multi-stage fusion procedures [9]- [15].…”

mentioning

confidence: 99%

See 4 more Smart Citations

Distributed Dynamic Map Fusion via Federated Learning for Intelligent Networked Vehicles

Zhang,

Wang,

Hong

et al. 2021

Preprint

View full text Add to dashboard Cite

The technology of dynamic map fusion among networked vehicles has been developed to enlarge sensing ranges and improve sensing accuracies for individual vehicles. This paper proposes a federated learning (FL) based dynamic map fusion framework to achieve high map quality despite unknown numbers of objects in fields of view (FoVs), various sensing and model uncertainties, and missing data labels for online learning. The novelty of this work is threefold: (1) developing a three-stage fusion scheme to predict the number of objects effectively and to fuse multiple local maps with fidelity scores; (2) developing an FL algorithm which fine-tunes feature models (i.e., representation learning networks for feature extraction) distributively by aggregating model parameters; (3) developing a knowledge distillation method to generate FL training labels when data labels are unavailable. The proposed framework is implemented in the Car Learning to Act (CARLA) simulation platform. Extensive experimental results are provided to verify the superior performance and robustness of the developed map fusion and FL schemes. I. INTRODUCTIONThe intelligent networked vehicle system (INVS) is an emerging vehicle-edge-cloud system that accomplishes cooperative perception, map management, planning and maneuvering tasks via vehicle-to-everything (V2X) communication [1]-[3]. Among all the tasks, distributed map management aims to enlarge sensing ranges and improve sensing accuracies for individual vehicles and plays a central role in . While static maps describe stationary objects (e.g., roads, buildings, and trees), dynamic maps emphasize updating information of mobile objects (e.g., pedestrians, cars, and animals) in real time. Fig. 1 shows the architecture of an intelligent networked vehicle system. There are three main steps for dynamic map fusion: (1) local sensing and perception, (2) local map fusion and uploading, and (3) global map fusion and broadcasting.

show abstract

mentioning

confidence: 99%