Autonomous driving challenge: To Infer the property of a dynamic object based on its motion pattern using recurrent neural network

Fathollahi, Mona; Kasturi, Rangachar

doi:10.48550/arxiv.1609.00361

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The studies (Alsharif et al, 2016;Fathollahi & Kasturi, 2016) use ANN to discuss IoT device connection and wireless communication. In this case, ANN is critical for improving driver behavior modeling (Morton et al, 2017), categorizing things (Fathollahi & Kasturi, 2016;Rausch et al, 2017), and predicting mobility speed (Alsharif et al, 2016). The IoT also covers the growing popularity of entities and things that use unique IDs to autonomously send data over a network.…”

Section: Connectivity and Qosmentioning

confidence: 99%

Evolution toward intelligent communications: Impact of deep learning applications on the future of 6G technology

Abd Elaziz,

Al‐qaness,

Dahou

et al. 2023

WIREs Data Min & Knowl

View full text Add to dashboard Cite

The sixth generation (6G) represents the next evolution in wireless communication technology and is currently under research and development. It is expected to deliver faster speeds, reduced latency, and greater capacity compared to the current 5G wireless technology. 6G is envisioned as a technology capable of establishing a fully data‐driven network, proficient in analyzing and optimizing end‐to‐end behavior and handling massive volumes of real‐time data at rates of up to terabits per second (Tb/s). Moreover, 6G is designed to accommodate an average of 1000+ substantial connections per person over the course of a decade. The concept of a data‐driven network introduces a new service paradigm, which offers fresh opportunities for applications within 6G wireless communication and network design in the future. This paper aims to provide a survey of existing applications of 6G that are based on deep learning techniques. It also explores the potential, essential technologies, scenarios, challenges, and related topics associated with 6G. These aspects are crucial for meeting the requirements for the development of future intelligent networks. Furthermore, this work delves into various research gaps between deep learning and 6G that remain unexplored. Different potential deep learning applications for 6G networks, including privacy, security, environmentally friendly communication, sustainability, and various wireless applications, are discussed. Additionally, we shed light on the challenges and future trends in this field.This article is categorized under: Technologies > Computational Intelligence Fundamental Concepts of Data and Knowledge > Explainable AI Technologies > Machine Learning

show abstract

Section: Connectivity and Qosmentioning

confidence: 99%

Evolution toward intelligent communications: Impact of deep learning applications on the future of 6G technology

Abd Elaziz,

Al‐qaness,

Dahou

et al. 2023

WIREs Data Min & Knowl

View full text Add to dashboard Cite

show abstract

“…In turn, these tracking approaches provide the tools to collect motion pattern information of surronding dynamic obstacles such that this information may help to classify obstacles depending on their dynamic properties [53].…”

Section: ) Traversable Area Segmentationmentioning

confidence: 99%

Self-supervised learning for autonomous vehicles perception: A conciliation between analytical and learning methods

Chiaroni,

Rahal,

Hueber

et al. 2019

Preprint

View full text Add to dashboard Cite

T HE interest for autonomous driving has continuously increased during the last two decades. However, to be adopted, such critical systems need to be safe. Concerning the perception of the ego-vehicle environment, the litterature has investigated two different types of methods. On the one hand, analytical methods, also referred to as hand-crafted, are generally designed from end-to-end. On the other hand, learning methods aim to design their proper representation of the observed scene.Analytical methods have demonstrated their usefulness for several tasks, including the keypoints detection [1], [2], optical flow, depth map estimation, background subtraction, geometric shape detection, tracking filtering, and simultaneous localization and mapping (SLAM) [3]. Those methods have the advantage to be explainable from end-to-end. However, it is difficult to apply them on high dimensional data for semantic scene analysis. For example, identifying the other users present in an urban scene requires to extract complex patterns from high dimensional data captured by camera sensors.Learning methods are nowadays the most adapted in terms of prediction performances for complex pattern recognition tasks [4] implied in autonomous vehicles scene analysis and understanding. However, the state-of-the-art results are often obtained with large and fully labeled training datasets [5]. Hand-labeling a large dataset for a given specific application has a cost. Another difficulty is to aprehend from end-to-end the learned representations. To overcome the former limitation, transfer learning and weakly supervised learning methods have appeared. Some of them can exploit partially labeled datasets [6], [7], or noisy labeled datasets [8], [9]. Concerning the latter problem, under mild theoretical assumptions on the learning model, we can interpret the predicted outputs. For instance, it is possible to automatically detect the training overfitting [10], to estimate the fraction of mislabeled examples [11], or estimate the uncertainty in the prediction outputs [12].Another challenge is to prevent unpredictable events. Indeed, some scenes unseen during the training can appear frequently in the context of the autonomous vehicle. For instance, an accident on the road can change drastically the appearance and the location of potential obstacles.VEDECOM Institute, Department of delegated driving (VEH08), Perception team,

show abstract

Autonomous driving challenge: To Infer the property of a dynamic object based on its motion pattern using recurrent neural network

Cited by 2 publications

References 0 publications

Evolution toward intelligent communications: Impact of deep learning applications on the future of 6G technology

Evolution toward intelligent communications: Impact of deep learning applications on the future of 6G technology

Self-supervised learning for autonomous vehicles perception: A conciliation between analytical and learning methods

Contact Info

Product

Resources

About