Design and Implementation of Edge-Fog-Cloud System through HD Map Generation from LiDAR Data of Autonomous Vehicles

Lee, Junwon; Lee, Kieun; Yoo, Aelee; Moon, Chang-Joo

doi:10.3390/electronics9122084

Cited by 23 publications

(12 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their neural network predicted the map elements observed in a bird's-eye view into a vectorized map. Meanwhile, Junwon et al [27] introduced the Edge-Fog-Cloud architecture to generate HD maps from vehicle sensor data. The Fog server performed the normal distribution transform-simultaneous localization and mapping (NDT-SLAM) to generate an HD map from the LiDAR point cloud, while the cloud server converted the generated HD map to a global coordinate frame for compilation.…”

Section: Automated Hd Map Generationmentioning

confidence: 99%

Lane-Level Map Generation and Management Framework Using Connected Car Data

Kim,

Moon,

Moon

2023

Electronics

Self Cite

View full text Add to dashboard Cite

This study proposes a lane-level map generation and management framework using connected sensor data to reduce the manpower and time required for producing and updating high-definition (HD) maps. Unlike previous studies that relied on the onboard processing capabilities of vehicles to collect map-constructing elements, this study offloads computing for map generation to the cloud, assigning vehicles solely the role of transmitting sensor data. For efficient data collection, we divide the space into a grid format to define it as a partial map and establish the state of each map and its transition conditions. Lastly, tailored to the characteristics of the road elements composing the map, we propose an automated map generation technique and method for selectively collecting data. The map generation method was tested using data collected from actual vehicles. By transmitting images with an average size of 350 KB, implementation was feasible even with the current 5G upload bandwidth. By utilizing 12,545 elements, we were able to achieve a position accuracy and regression RMSE of less than 0.25 m, obtaining 651 map elements to construct the map. We anticipate that this study will help reduce the manpower and time needed for deploying and updating HD maps.

show abstract

Section: Automated Hd Map Generationmentioning

confidence: 99%

Lane-Level Map Generation and Management Framework Using Connected Car Data

Kim,

Moon,

Moon

2023

Electronics

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, the database does not directly store and process point cloud data about buildings but only stores two-dimensional (2D) information in a top-view format, making it challenging to apply to the dynamic autonomous driving environments considered in the present study. Lee et al proposed a fog computing server model that applied the normal distributions transform-simultaneous localization and mapping (NDT-SLAM) technique to LiDAR data generated from vehicles to create an HD (high-definition) map [12]. This study was significant, as it used Hadoop to store large amounts of point cloud data and created precision maps using SLAM techniques.…”

Section: Point Cloud and Spatial Databasementioning

confidence: 99%

The Distributed HTAP Architecture for Real-Time Analysis and Updating of Point Cloud Data

Kim,

Moon

2023

Electronics

Self Cite

View full text Add to dashboard Cite

Updating the most recent set of point cloud data is critical in autonomous driving environments. However, existing systems for point cloud data management often fail to ensure real-time updates or encounter situations in which data cannot be effectively refreshed. To address these challenges, this study proposes a distributed hybrid transactional/analytical processing architecture designed for the efficient management and real-time processing of point cloud data. The proposed architecture leverages both columnar and row-based tables, enabling it to handle the substantial workloads associated with its hybrid architecture. The construction of this architecture as a distributed database cluster ensures real-time online analytical process query performance through query parallelization. A dissimilarity analysis algorithm for point cloud data, built by utilizing the capabilities of the spatial database, updates the point cloud data for the relevant area whenever the online analytical process query results indicate high dissimilarity. This research contributes to ensuring real-time hybrid transactional/analytical processing workload processing in dynamic road environments, helping autonomous vehicles generate safe, optimized routes.

show abstract

“…Connected Vehicles: In future autonomous vehicles, an object detection task can be performed collaboratively using distributed computing or joint training and inference [18], [20]. Methods proposed in this category include deploying AI models through edge computing, fog computing, cloud computing or their respective combinations [3], [22], [31], [33], [36]. As the AI model, such as DNN, can be dense in size and may require high computational resources, it is practically challenging to deploy them on Edge devices.…”

Section: A Object Detection Using Edge In Autonomous Vehiclesmentioning

confidence: 99%

Bias Detection and Generalization in AI Algorithms on Edge for Autonomous Driving

Katare

Kourtellis

Park

et al. 2022

2022 IEEE/ACM 7th Symposium on Edge Computing (SEC)

View full text Add to dashboard Cite

A machine learning model can often produce biased outputs for a familiar group or similar sets of classes during inference over an unknown dataset. The generalization of neural networks have been studied to resolve biases, which has also shown improvement in accuracy and performance metrics, such as precision and recall, and refining the dataset's validation set. Data distribution and instances included in test and validation-set play a significant role in improving the generalization of neural networks. For producing an unbiased AI model, it should not only be trained to achieve high accuracy and minimize false positives. The goal should be to prevent the dominance of one class/feature over the other class/feature while calculating weights. This paper investigates state-of-art object detection/classification on AI models using metrics such as selectivity score and cosine similarity. We focus on perception tasks for vehicular edge scenarios, which generally include collaborative tasks and model updates based on weights. The analysis is performed using cases that include the difference in data diversity, the viewpoint of the input class and combinations. Our results show the potential of using cosine similarity, selectivity score and invariance for measuring the training bias, which sheds light on developing unbiased AI models for future vehicular edge services.

show abstract

Design and Implementation of Edge-Fog-Cloud System through HD Map Generation from LiDAR Data of Autonomous Vehicles

Cited by 23 publications

References 9 publications

Lane-Level Map Generation and Management Framework Using Connected Car Data

Lane-Level Map Generation and Management Framework Using Connected Car Data

The Distributed HTAP Architecture for Real-Time Analysis and Updating of Point Cloud Data

Bias Detection and Generalization in AI Algorithms on Edge for Autonomous Driving

Contact Info

Product

Resources

About