Acoustic waves propagate slowly in water, and time-varying UACs result in inevitably high bit error rate and packet loss rate. The long propagation delay and the error-prone nature of UACs impose challenges on reliable transmissions in UANs. In this article, we identify the challenges for reliable acoustic transmissions and propose a CL-FEC scheme, which achieves opportunistic transmissions to overcome the frequent transmission failures in UACs. CL-FEC adopts fountain codes as a packet-level FEC and adopts channel codes as a bit-level FEC, to realize reliable transmissions over UACs without per-packet feedback. To further improve the throughput of CL-FEC, we formulate the transmissions over UACs into a stochastic throughput optimization problem. A discrete stochastic approximation based algorithm is then developed to achieve the optimal CL-FEC by online exploiting channel estimating and algorithm iterations. Simulation results show the asymptotic convergence and the iterative optimality of the algorithm.

中文翻译：

---

水下声网的可靠机会传输

声波在水中传播缓慢，随时间变化的UAC不可避免地会导致较高的误码率和丢包率。UAC的长传播延迟和容易出错的性质给UAN中的可靠传输带来了挑战。在本文中，我们确定了可靠的声音传输所面临的挑战，并提出了一种CL-FEC方案，该方案实现了机会传输，从而克服了UAC中频繁发生的传输故障。CL-FEC采用源代码作为分组级FEC，并采用信道代码作为比特级FEC，以实现在UAC上的可靠传输而无需逐包反馈。为了进一步提高CL-FEC的吞吐量，我们将UAC上的传输公式化为随机吞吐量优化问题。然后，开发了一种基于离散随机逼近的算法，通过在线利用信道估计和算法迭代来获得最佳的CL-FEC。仿真结果表明了算法的渐近收敛性和迭代最优性。