A Centralized Discovery-Based Method for Integrating Data Distribution Service and Time-Sensitive Networking in In-Vehicle Networks

Feng Luo,Yi Ren,Yanhua Yu,Yunpeng Li,Zitong Wang

2024-09-06

Abstract:As the electronic and electrical architecture (E/EA) of intelligent and connected vehicles (ICVs) evolves, traditional distributed and signal-oriented architectures are being replaced by centralized, service-oriented architectures (SOA). This new generation of E/EA demands in-vehicle networks (IVNs) that offer high bandwidth, real-time, reliability, and service-oriented. data distribution service (DDS) and time-sensitive networking (TSN) are increasingly adopted to address these requirements. However, research on the integrated deployment of DDS and TSN in automotive applications is still in its infancy. This paper presents a DDS over TSN (DoT) communication architecture based on the centralized discovery architecture (CDA). First, a lightweight DDS implementation (FastDDS-lw) is developed for resource-constrained in-vehicle devices. Next, a DDS flow identification algorithm (DFIA) based on the CDA is introduced to identify potential DDS flows during the discovery phase automatically. Finally, the DoT communication architecture is designed, incorporating FastDDS-lw and DFIA. Experimental results show that the DoT architecture significantly reduces end-to-end latency and jitter for critical DDS flows compared to traditional Ethernet. Additionally, DoT provides an automated network configuration method that completes within a few tens of milliseconds.

Networking and Internet Architecture,Systems and Control

What problem does this paper attempt to address?

This paper attempts to address the problem of how to effectively integrate Data Distribution Service (DDS) and Time-Sensitive Networking (TSN) in Intelligent Connected Vehicles (ICVs) to meet the In-Vehicle Networks (IVNs) requirements for high bandwidth, real-time performance, reliability, and service orientation. Specifically, the paper focuses on the following challenges:

1. **Challenges at the DDS level**:
- The increased complexity of DDS software leads to high resource consumption and uncontrollable software processing delays.

2. **Challenges at the TSN level**:
- Due to service updates and upgrades, DDS traffic dynamically changes, and there is a lack of automated methods to identify DDS traffic and update TSN configurations.

To address these challenges, the paper proposes the following contributions:

1. **Lightweight DDS software**:
- Implemented a lightweight DDS (FastDDS-lw) that includes only the QoS policies required by IVNs, reducing the complexity of DDS software.

2. **DDS traffic identification algorithm**:
- Automatically identifies changing DDS traffic based on a Centralized Discovery Architecture (CDA).

3. **Communication architecture DoT**:
- Designed a DoT communication architecture that integrates DDS and TSN and supports automated network configuration in IVNs.

Through these contributions, the paper aims to significantly reduce the end-to-end delay and jitter of critical DDS traffic and provide an automated network configuration method that completes within tens of milliseconds.