The evaluation of geometry-based stochastic models for device-to-device channels

Device-to-device (D2D) communication is becoming a critical technology to improve the spectral efficiency in cellular communication systems. Although many upper-layer protocols have been designed for D2D communications, these work did not consider the specific characteristics of D2D channels, such as low antenna height and dual mobility, due to the lack of adequate D2D channel models. To promote research on these aspects, in this paper, different geometry-based stochastic models (GBSMs) are evaluated to facilitate channel-oriented D2D communication system design and optimization. Firstly, based on the specific characteristics of D2D channels, different GBSMs are introduced for D2D channels, their spatial correlation and Doppler properties are analyzed. Next, the space-time-frequency characteristics metrics including spatial cross-correlation function and Doppler power spectral density are used to evaluate the different channel models with different scenarios. Simulation results show that GBSMs can be easily modified to adapt different D2D communication scenarios, and the application of GBSMs have enormous relationship with D2D communication scenarios.