Optimizing the Transmission Range in an Underwater Acoustic Network Arnau Porto and Milica Stojanovic Massachusetts Institute of Technology Cambridge, MA, USA E-mail: porto@mit.edu, millitsa@mit.edu Abstract—An extension of Distance-Aware Collision Avoidance Protocol (DACAP) [1] is proposed that permits its implementation in large networks where maximal connectivity is not available. The technique proposed increases the energy efficiency by opti- mizing the transmission power of the nodes. The optimal power is the minimal power that still guarantees connectivity between each node and the sink. Simulation results show that this transmission power also results in throughput maximization. For a network of nodes uniformly distributed within a rectangular grid, the optimal transmission range can be determined as a function of the node density. A closed form approximation for this dependence is obtained. Index Terms— MAC, power control, optimal transmission range, underwater acoustic networks, energy efficiency, through- put. I. I NTRODUCTION Underwater acoustic networks present several challenges for the design of medium access control (MAC) protocols. First, latency is very high because of the low speed of sound (nominally 1500 m/s). Additionally, attenuation depends both on the distance l and the frequency of the signal f , as A(l, f )= l k · a(f ) l (1) where k is the spreading factor and a(f ) is the absorption coefficient, which increases with frequency. As a consequence, bandwidth is limited, and it decreases with transmission dis- tance [2]. Finally, because the network lifetime depends on the batteries of the nodes, minimizing energy consumption is an important design goal. In this paper we are focusing on random channel access, which is useful in scenarios with a large number of nodes, which transmit in a bursty manner at a relatively low duty cy- cle. Aloha is the predecessor and the basis of many contention oriented protocols [3]. However, its high collision rate requires too much power in underwater acoustic systems [1]. At the same time, because of the high latency, carrier sensing is not reliable. For this reason, MACA (Multiple Access with Col- lision Avoidance) [4] type of protocols have been considered for underwater networks. MACA is based on an exchange of RTS/CTS (request to send / clear to send) commands that se- cure the communication channel. This protocol has been used in an experimental underwater network [5]. Current research is focusing on MACA type protocols such as S-FAMA (Slotted- Floor Acquisition Multiple Access) [6], PCAP (Propagation- delay-tolerant Collision-Avoidance Protocol) [7] and DACAP (Distance-Aware Collision Avoidance Protocol) [1]. Another approach considered for channel sharing in emergy-limited 0 1.25 2.5 3.75 5 0 1.25 2.5 3.75 5 Fig. 1. Simulation scenario: 25 km 2 , 16 relays (circles), three AUVs (triangles with associated trajectories) and one sink (square in the middle). underwater acoustic networks is that of sleeping schedules [8], [9]. DACAP offers good throughput performance, and, unlike S-FAMA, it is suitable for asynchronous nodes. DACAP was designed for applications with and without acknowledgments. However, it was only analyzed for maximally connected networks with fixed nodes. Maximal connectivity is achieved when any node can communicate with any other node via a single hop. Our goal is to extend DACAP to situations without maximal connectivity, so as to make it scalable to networks with large coverage. In doing so, we aim to optimize the transmission power so as to minimize the overall power consumption. Optimization of the transmission range was considered in [10] for terrestrial radio networks; however, no similar analyses are available for underwater acoustic networks. The paper is organized as follows. In Sec.II, DACAP is overviewed. In Sec.III, optimization of the transmission range is discussed and simulation results are presented. Sec.IV. establishes the relationship between transmission range and node density for a network with uniform node distribution.