Adaptive Probabilistic Medium Access in MPR-Capable Ad-hoc Wireless Networks Majid Ghanbarinejad Department of Electrical and Computer Engineering University of Alberta Edmonton, Canada Email: madjid@ece.ualberta.ca Christian Schlegel and Pawel Gburzynski Department of Computing Science University of Alberta Edmonton, Canada Email: {schlegel, pawel}@cs.ualberta.ca Abstract—Medium access in ad-hoc wireless networks must be performed in a distributed fashion due to lack of coordination between nodes. Specifically, when nodes are capable of receiving more than one transmission simultaneously, the design of dis- tributed medium-access mechanisms that efficiently exploit the receiver’s capability becomes more challenging. Adaptive probabilistic medium access for ad-hoc wireless networks is proposed in this paper. Nodes with data packets to transmit perform an announcement process in order to inform other nodes of their intended traffic. The acquired information through this process about other potential transmitters in the vicinity is then used by the nodes to choose a transmission probability with which they transmit their data packets. The performance of a multi-packet reception capable ad-hoc wireless network under the proposed protocol is analyzed and evaluated numerically and via simulations, and compared with Aloha-type random access. I. I NTRODUCTION After successful realizations of infrastructure wireless net- works in the fields of cellular communications and local area networking, infrastructure-less networking is going to play a key role in the next generation of wireless networks. Applications of such networks, also known as ad-hoc wireless networks, include networking wireless sensors in industrial environments, surveillance, disaster recovery, and distributed computing. The fundamental differences between these net- works and their traditional infrastructure counterparts raise new challenges. Examples are distributed algorithms for rout- ing, medium access, quality of service, and security [1], low- power implementations, and information theoretic evaluations of such networks [2]. Multi-packet reception (MPR), powered by new technolo- gies such as code-division multiple-access, multiple-antenna arrays, and space-time coding, is the ability to receive multiple communications packets simultaneously. The ability of receiv- ing concurrent transmissions deviates from the traditional col- lision model and, hence, makes motivation towards designing medium access mechanisms to be able to exploit this feature. Due to lack of central coordination, medium access in ad-hoc wireless networks must be controlled by distributed mechanisms. Specifically, when nodes are capable of receiving concurrent transmissions, they need distributed protocols to ex- ploit the capability optimally in the the sense of using available bandwidth while avoiding excessive concurrent transmissions. To approach this goal, we propose adaptive probabilistic medium access through which a potential transmitter acquires an estimate of the number of on-going transmissions in its vicinity and, then, transmits its packet with a probability adapted to the current capacity of the receiver. In this method, having more potential transmitters in an area leads to lower transmission probabilities. Finding the optimum transmission probability leads to an optimization problem with regard to overall network performance. In this paper, the performance of an MPR-capable ad-hoc wireless network under the proposed protocol is studied. The network performance is analyzed and evaluated numerically and via simulations, and compared with Aloha-type random access. Hence, the scope of this paper is not to propose a detailed medium-access control (MAC) protocol, but to introduce a feasible medium-access mechanism followed by an analytical framework. The rest of the paper is organized as follows. In Section II, the related literature is reviewed. In Section III, the system model is introduced and the proposed protocol is described. The network performance is analyzed in Section IV. Numerical evaluations and simulation results are provided in Section V. Finally, Section VI concludes the paper. II. RELATED LITERATURE MPR-capable cellular wireless networks have been studied and several MAC protocols have been proposed in the liter- ature [3]–[6]. The protocol proposed in [3] utilizes explicit reservations. In this method, time is divided into slots, each composed of two sub-slots: reservation and data transmission. Nodes with packets to transmit enter the reservation phase that deterministically specifies the winners. Similarly, protocols proposed in [4] and [5] require central controllers at the base stations. A predictive protocol utilizing a finite-size buffer was proposed in [7]. The performance of the protocol and the effect of the buffer was analyzed using the theory of discrete- time Markov chains. The analysis, however, is computationally complex and was offered as a complement to simulations. In [6], a multi-group priority queuing MAC protocol for cellular wireless networks was proposed. The protocol performs a