-
DVS-RG: Differential Variable Speed Limits Control using Deep Reinforcement Learning with Graph State Representation
Authors:
**gwen Yang,
** Wang,
Fatemeh Golpayegani,
Shen Wang
Abstract:
Variable speed limit (VSL) control is an established yet challenging problem to improve freeway traffic mobility and alleviate bottlenecks by customizing speed limits at proper locations based on traffic conditions. Recent advances in deep reinforcement learning (DRL) have shown promising results in solving VSL control problems by interacting with sophisticated environments. However, the modeling…
▽ More
Variable speed limit (VSL) control is an established yet challenging problem to improve freeway traffic mobility and alleviate bottlenecks by customizing speed limits at proper locations based on traffic conditions. Recent advances in deep reinforcement learning (DRL) have shown promising results in solving VSL control problems by interacting with sophisticated environments. However, the modeling of these methods ignores the inherent graph structure of the traffic state which can be a key factor for more efficient VSL control. Graph structure can not only capture the static spatial feature but also the dynamic temporal features of traffic. Therefore, we propose the DVS-RG: DRL-based differential variable speed limit controller with graph state representation. DVS-RG provides distinct speed limits per lane in different locations dynamically. The road network topology and traffic information(e.g., occupancy, speed) are integrated as the state space of DVS-RG so that the spatial features can be learned. The normalization reward which combines efficiency and safety is used to train the VSL controller to avoid excessive inefficiencies or low safety. The results obtained from the simulation study on SUMO show that DRL-RG achieves higher traffic efficiency (the average waiting time reduced to 68.44\%) and improves the safety measures (the number of potential collision reduced by 15.93\% ) compared to state-of-the-art DRL methods.
△ Less
Submitted 15 May, 2024;
originally announced May 2024.
-
An Adaptive System Architecture for Multimodal Intelligent Transportation Systems
Authors:
Muhammad Farooq,
Nima Afraz,
Fatemeh Golpayegani
Abstract:
Multimodal intelligent transportation systems (M-ITS) encompass a range of transportation services that utilise various modes of transport and incorporate intelligent technologies for enhanced efficiency and user experience. There are several challenges in M-ITS including data integration, Interoperability, scalability, user experience, etc. To address these challenges, such a system requires an a…
▽ More
Multimodal intelligent transportation systems (M-ITS) encompass a range of transportation services that utilise various modes of transport and incorporate intelligent technologies for enhanced efficiency and user experience. There are several challenges in M-ITS including data integration, Interoperability, scalability, user experience, etc. To address these challenges, such a system requires an adaptive system architecture that enables M-ITS to operate as an integrated ecosystem. In this paper, we provide an adaptive, user-centric, and layered architecture for multimodal transportation systems. The proposed architecture ensures scalability for seamless interactions of various subcomponents, that are often managed by different stakeholders. Concurrently, the data architecture is detailed, covering diverse data sources, advanced analytics, and stringent governance, providing a robust basis for intelligent decision-making. We provide two example use cases of the proposed architecture, showing how the data architecture and the system architecture can be fused and serve multimodal intelligent transport services.
△ Less
Submitted 13 February, 2024;
originally announced February 2024.
-
On the Sum Secrecy Rate Maximisation for Wireless Vehicular Networks
Authors:
Muhammad Farooq,
Le-Nam Tran,
Fatemeh Golpayegani,
Nima Afraz
Abstract:
Wireless communications form the backbone of future vehicular networks, playing a critical role in applications ranging from traffic control to vehicular road safety. However, the dynamic structure of these networks creates security vulnerabilities, making security considerations an integral part of network design. We address these security concerns from a physical layer security aspect by investi…
▽ More
Wireless communications form the backbone of future vehicular networks, playing a critical role in applications ranging from traffic control to vehicular road safety. However, the dynamic structure of these networks creates security vulnerabilities, making security considerations an integral part of network design. We address these security concerns from a physical layer security aspect by investigating achievable secrecy rates in wireless vehicular networks. Specifically, we aim to maximize the sum secrecy rate from all vehicular pairs subject to bandwidth and power resource constraints. For the considered problem, we first propose a solution based on the successive convex approximation (SCA) method, which has not been applied in this context before. To further reduce the complexity of the SCA-based method, we also propose a low-complexity solution based on a fast iterative shrinkage-thresholding algorithm (FISTA). Our simulation results for SCA and FISTA show a trade-off between convergence and runtime. While the SCA method achieves better convergence, the FISTA-based approach is at least 300 times faster than the SCA method.
△ Less
Submitted 30 January, 2024;
originally announced January 2024.
