Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
In collaboration with Vanderbilt University, NIST researchers have published a paper, “Scalable HLA Co-Simulations of Connected and Automated Vehicles using Aggregation of Virtual Federates,” on improving the scalability of network simulations. This paper, which was presented at the 2024 IEEE Workshop on Design Automation for CPS and IoT (DESTION), describes a method to improve the performance of co-simulations by aggregating multiple controllers into a single process.
A co-simulation integrates multiple software processes and physical hardware into a cooperative simulation to create high-fidelity system models. While co-simulations can produce more accurate models of complex systems than standalone simulators, they tend to have poor performance and major scalability challenges when many processes are integrated into the cooperative simulation. This is a significant problem for Internet of Things (IoT) systems which contain hundreds to thousands of independent devices that communicate with each other to control the system.
This paper introduces the concept of an aggregator that represents multiple IoT devices and integrates the devices as a single process into the co-simulation. The aggregation method was implemented to maintain unique identifiers for each device allowing them to directly interact with each other and other processes in the co-simulation, unaware that their messages are being aggregated. A co-simulation of connected and automated vehicles (CAVs), where road vehicles communicate with each other to influence their driving behavior using vehicle-to-vehicle (V2V) communications, was implemented to demonstrate the performance benefits of the aggregation method. A simple road network containing up to 36 communicating vehicles was simulated both with and without using the new aggregation method. The use of aggregation resulted in, on average, a 15 % reduction to execution time and up to a 95 % reduction in the memory requirements to run the co-simulation.