Abstract:In the autonomous driving era, high-definition (HD) maps are an essential building block to enable fine-grained environmental perception, precise localization, and path planning. However, with rich multidimensional information, the size of HD map data is huge and cannot be stored onboard, where the dynamic map data need to be distributed in real time via vehicular networks and how to design the distribution mechanism (i.e., determining the data source for requests) becomes crucial. For the end-to-end communication protocols (i.e., TCP/IP), the main limitation is the vehicle mobility and high dynamic of the network topology, which can degrade the transmission performance dramatically. Therefore, in this article, we propose a reinforcement learning-based data source selection scheme, named RLSS, for efficient HD map distribution in vehicular named data networking (NDN) scenarios, which aims at seeking the best data source (roadside infrastructures or nearby vehicles) in accordance with the map data requests. Specifically, in RLSS, we adopt a deep reinforcement learning-based architecture to learn a neural network as an agent to make the decision of data source selection, which can work online after offline training based on historical selection action performance. In addition, to solve the “cold start” problem for a new vehicle, we propose a model aggregation algorithm and weight update approach to learn the model parameters from its nearby vehicles, which can guarantee the performance while saving the communication cost. Finally, we implement RLSS in NS-3 by adopting the tools of the ndnSIM, SUMO, and Gym. Extensive simulations demonstrate that RLSS can significantly improve the transmission performance in terms of delay, throughput, and packet loss when compared with state-of-the-art data source selection schemes.

High Definition Map Distribution in Named Data Networking Based VANETs

An Energy-Efficient High Definition Map Data Distribution Mechanism for Autonomous Driving

NDNVIC: Named Data Networking for Vehicle Infrastructure Cooperation.

RLSS: A Reinforcement Learning Scheme for HD Map Data Source Selection in Vehicular NDN

Enhancement of High-definition Map Update Service Through Coverage-aware and Reinforcement Learning

A scalable, dynamic, and secure traffic management system for vehicular named data networking applications

NIHR: Name/ID Hybrid Routing in Information-centric VANET.

Neural HD Map Generation from Multiple Vectorized Tiles Locally Produced by Autonomous Vehicles

High-Definition Map Generation Technologies For Autonomous Driving

Boosting named data networking for data dissemination in urban VANET scenarios

Love of Variety Based Latency Analysis for High Definition Map Updating: Age of Information and Distributional Robust Perspectives

Creation of High Definition Map for Autonomous Driving Within Specific Scene

Review and Challenge: High Definition Map Technology for Intelligent Connected Vehicle

High Definition Map for Automated Driving: Overview and Analysis

Open HD map service model: an interoperable high-Definition map data model for autonomous driving

Data dissemination scheme based on adaptive density of nodes in vehicular ad-hoc networks

VI-Map: Infrastructure-Assisted Real-Time HD Mapping for Autonomous Driving

SDN-Based Resource Management for Autonomous Vehicular Networks: A Multi-Access Edge Computing Approach

Design of V2X-Based Vehicular Contents Centric Networks for Autonomous Driving

Optimizing QoS in HD Map Updates: Cross-Layer Multi-Agent with Hierarchical and Independent Learning

A Review on End-to-End High-Definition Map Generation