top of page

Resilient V2X Communication for Cooperative and Remote Driving 

Project Description: Recent years have witnessed autonomous driving making successful strides. The remaining challenges stopping wide adoption mostly lie in the long tail, handling corner cases. These corner cases are difficult partly due to a) the limited visibility of ego-vehicles due to obstructions and other conditions, despite a full suite of sensors; and b) complex or unforeseen scenarios that an AV is not designed for. These two challenges have led to the notion of cooperative driving, where connected vehicles share sensor data and other information, and remote driving, where a remote human operator takes over control of an AV when needed.  Both of these require stable and secure V2V and V2I connection; stringent latency and high bandwidth (especially, the uplink) for video and LiDAR delivery and sharing. This project will build end-to-end cooperative and remote driving pipelines with C-V2X radios and infrastructure connections to build the foundation for a resilience study.  
 

  • Stabilizing V2I Connection for Remote Driving. Vehicles stream a massive amount of onboard sensor data to remote operators who fully immerse themselves in the situation to make driving decisions. Such a connection needs extreme stability with low latency and low jitter. It is challenging in vehicular dynamics where cellular tower handovers could easily interrupt these connections. Leveraging vehicle sensors, there is an opportunity to predict vehicle trajectory and proactively provision resources to perform predictive handover.  

  • Opportunistic V2V communications to compensate cellular channel dynamics. While using sensors and measurements we can identify and predictively handover cellular connections, there are still channel dynamics that hinder stable streaming. When this happens, V2V direct communication can offer a backup relay to ensure the remote operator won’t lose control in critical moments. 

  • Cooperative autonomy using V2X communication. Cooperative driving enables vehicles to share sensor data over V2X connections to help augment each individual vehicle’s perspective. Leveraging this augmented view, onboard autonomy can be improved substantially for both complete independent self-driving and the local operation during a remote session. 

  • PNT Co-optimization using both Cellular and GNSS Signals. 5G V2X communication needs precise timing as a premise to establish synchronized transmissions. Similar to GNSS signals, cellular signals also carry timing information. Surveyed cellar stations can also function as RTK base stations. Integrating cellular signals into PNT systems can offer great potential for joint optimization, benefiting both V2X synchronization as well as GNSS time-of-flight calculation.  

​

US DOT Priorities:  This research project directly targets the US DOT’s research priority area of Reducing Transportation Cybersecurity Risks. Specifically, we will be investigating potential risks of safety-critical applications over opportunistic V2X connections under spoofing attacks. 

Outputs:  In this project, we will: 

  • Task 1: Build multi-vehicle testbed, collect dataset to measure V2X and cellular channel characteristics, and identify challenges and opportunities for remote and cooperative driving. 

  • Collect raw V2X channel measurement data using commercial V2X radios 

  • Analyze channel resource allocation, interference, and packet collisions under high data rate requests from multi-agent scenarios. 

  • Collect cellular connection data at a bigger scale (across cell towers, potentially in lower signal strength/coverage areas) 

  • Task 2: Investigate and develop new methods for predictive handover for seamless resilient V2I/cellular connections.  

  • Leverage mapping and trajectory plans to predict potential handover scenarios 

  • Provision the connection ahead of time, potentially using V2V replays to bridge the gap between handovers 

  • Task 3: Investigate and develop new methods incorporating both cellular and GNSS to enhance the resilience and accuracy of the timing system.  

  • Collect raw GNSS data and cellular metadata  

  • Analyze the timing signals from both, cross-check their accuracy and consistency 

  • Jointly optimize timing and synchronization 

​

We expect interest in this research from Connected and Autonomous Vehicle (CAV) navigation system designers, CAV application designers, and CAV manufacturers. We will actively encourage those on CARNATIONS External Advisory Board to contribute feedback and collaborate throughout the effort.  

Outputs/Impacts: CAVs offer a unique potential to address negative aspects of transportation systems (e.g., congestion, emissions) as well as enhance safety, if the navigation systems can continuously and reliably provide accurate vehicle state information at the bandwidth and sample rates required for vehicle control. The aim of this project is to rigorously analyze and demonstrate the extent to which the multiple sensors onboard CAVs can meet the application requirements.  

​

Final Research Report: Upon completion of the project we will provide a link to the final report. 

Image by Bernd 📷 Dittrich
bottom of page