Hierarchical Software Architectures and Vehicular Path Prediction for Cooperative Driving Applications

This paper investigates software architectures and path prediction techniques for vehicle-to-vehicle (V2V) cooperative systems. It proposes how software architectures should be constructed to enable the greatest flexibility for different cooperative driving applications while leveraging the full advantage of wireless communications. The applications may be designed for safety, mobility, or comfort objectives and incorporate autonomous behaviors available with non-communication-based sensing. The paper elaborates on the fundamental element of environment mapping, which includes vehicular path histories, vehicular path predictions, and target classifications. Within vehicular path prediction, the paper focuses on parametric and non-parametric path prediction approaches and illustrates the advantages and disadvantages of various existing methods.