LAUSR

Next-generation wireless networks are expected to support delay-sensitive applications such as augmented reality (AR) and virtual reality (VR). In AR/VR applications, computational jobs are offloaded to the edge computing server. Such offloading process requires transmissions of large data packets with low latency. Terahertz (THz) band can provide hundreds of GHz bandwidth and thus is promising in enabling ultra-high speed data transmissions. However, THz transmission distance is limited due to its inherently severe propagation loss. This paper considers to improve THz transmission distance with the help of mobile computing. The key idea is that, if mobile devices can handle partial computational jobs, the data transmission rate requirement is lowered, which in turn helps to improve THz transmission distance. Specifically, we aim to maximize THz transmission distance by jointly optimizing users' job offloading decisions and transmission carrier frequencies. To that end, we first introduce a delay-violation probability to characterize the delay constraints for delay-sensitive services. Based on the derived delay-violation probability, we formulate a THz transmission distance maximization (THz-TDM) problem. This THz-TDM problem is non-convex. To solve this issue, we derive in closed-form the expression of the achievable transmission distance with respect to carrier frequency and data rate. Subsequently, we propose a Hungarian-ADMM based iterative algorithm (HAB) and a low-complexity Rearrangement Inequality-like algorithm (RIL). The proposed RIL involves no iterations but only applies to the case where the edge computing server queue stability constraint is relaxed. Numerical results show that the developed schemes can effectively improve THz transmission distance.