True-Asynchronous Transmission Modeling and Software-Defined-Radio Experimental Testbed for Cognitive Radio Ad-Hoc Wireless Networks

The designs and analyses of "asynchronous" channel-hopping (CH) sequences were traditionally based on a "slot-synchronous" assumption. This assumption could not truly support the asynchronous and ad-hoc settings of cognitive-radio ad-hoc wireless networks (CRAHWNs) and resulted in overoptimistic performances. In this paper, a new "true-asynchronous" transmission model is studied, in which every time slot is divided into two or more equal sub-slots for rendezvous. All secondary users (SUs) do not need to synchronize in any level, not even the sub-slot level. Pairs of SUs can rendezvous under any arbitrary time shift between their CH sequences without requiring time-synchronization hardware/algorithms, coordination, or scheduling. The throughput and average time-to-rendezvous (ATTR) of CH sequences under the new model with fixed and adjustable payloads and different numbers of sub-slots are derived and compared with the conventional "slot-synchronous" transmission model. This paper also reports an experimental testbed using software-defined radios to validate the theoretical throughput and ATTR of the new transmission model. The numerical, experimental, and simulation studies find that the new transmission model with adjustable payload can support "true-asynchronous" operations and also achieve better ATTR than and the throughput upper bound of the conventional model without requiring time-synchronization hardware/algorithms or coordination.