Southwest Research Institute (SwRI) uses agent-based modeling (ABM) and simulation methods to gain insight into the behavior of small to large vehicle systems composed of individual vehicles, both manned or autonomous. This insight helps SwRI to design effective cooperative behaviors for intelligent vehicles and to understand how the introduction of a specific technology may impact the large-scale traffic system dynamics.
SwRI utilizes both commercially-available modeling and simulation tools, as well as custom-developed software tools, to create traffic system micro-simulations using an agent-based approach. The collective behavior of a traffic system emerges from the combined behavior of numerous self-directed vehicles; ABM is used to analyze this type of complex system. Each individual vehicle can be modeled in software as an “agent” of the system, and then simulated by instantiating a number of agents within a given set of environmental and behavioral limits.
Key Elements of SwRI Research
- High-fidelity traffic system micro-simulation for connected and cooperative vehicle systems
- Integration with commercially-available software tools
- Hardware-in-the-loop simulation capabilities, for devices such as DSRC radios, or for entire vehicle systems
SwRI has incorporated agent-based modeling and simulation since 2008 in the development of the MARTI® autonomous vehicle, for algorithm development and testing of behaviors such as intersection negotiation, cooperative platooning and cooperative sensor sharing. SwRI has also used this approach to examine the effects of specific communication technologies and cooperative vehicle algorithms on the behavior of larger traffic systems, involving hundreds or thousands of vehicles.