Lightweight constrained application protocol (CoAP) has emerged as a common communication protocol for resource-constrained equipment in distribution internet of things (IoT). CoAP introduces two modes for data transmission, i.e., non-confirmed mode for reducing transmission delay and confirmed mode for reducing packet-loss ratio, which can be dynamically selected to satisfy the service requirements. However, there are still some challenges in dynamic CoAP mode selection, including incomplete information and differentiated quality of service (QoS) requirements of distributed IoT services. In this paper, we propose a upper confidence bound (UCB)-based dynamic CoAP mode selection algorithm for data transmission to address these challenges. The simulation results validate that, compared with the fixed mode selection algorithm, the proposed algorithm can flexibly balance the tradeoff between packet-loss ratio and transmission delay as well as satisfy the differentiated QoS in distribution IoT.

轻量级受限应用协议 (CoAP) 已成为分布式物联网 (IoT) 中资源受限设备的通用通信协议。CoAP引入了两种数据传输模式，即减少传输延迟的非确认模式和减少丢包率的确认模式，可以动态选择满足业务需求。然而，动态CoAP模式选择仍然存在一些挑战，包括分布式物联网服务的信息不完整和服务质量（QoS）要求差异化。在本文中，我们提出了一种基于置信上限 (UCB) 的动态 CoAP 模式选择算法，用于数据传输以解决这些挑战。仿真结果验证，与固定模式选择算法相比，