A considerable percentage of Internet of Things end-devices are characterised by mobility, a feature that adds extra complexity to protocols used in Wireless Sensor Networks. LoRa is one of the newly introduced wireless sensor protocols, capable of delivering messages in long distances and consuming low energy, features that make it proper for low cost devices. Although LoRa was introduced as a technology for stationary devices, it can also be used for mobile devices of low speed. In this paper, we introduce an enhancement to Adaptive Data Rate (ADR) mechanism to enable mobile LoRa, by improving the connection reliability of mobile end-devices, while keeping energy consumption at low levels. Firstly, we propose the Linear Regression-ADR (LR-ADR) mechanism for the Network Server side to smooth the Signal to Noise Ratio (SNR) estimates per gateway and predict the SNR of the next transmission. Secondly, we propose the Linear Regression + ADR (LR+ADR) mechanism, an adaptive method for the end-device side to regain the connectivity faster with the Network Server. We conducted simulation modelling to evaluate the performance of our implementation while we compared our results with four alternative solutions ADR, ADR+, EMA-ADR, G-ADR. The results prove that our first approach (LR-ADR) performs better than the best competitor, and our second approach (LR+ADR) brings an additional improvement in terms of Packet Delivery Ratio (PDR), while they retain the Energy Consumption per Packet Delivered (ECPD) at low levels. In particular, in a scenario that mimics real world conditions, LR+ADR presents an increase of up to 520% for PDR compared to the original ADR and an improvement of up to 38% compared to the best competitor (G-ADR). Moreover, it reduces ECPD up to 74% compared to the original ADR, while keeping it at the same level with the best competitor (G-ADR).

相当大比例的物联网终端设备具有移动性，这一特性增加了无线传感器网络中使用的协议的额外复杂性。LoRa 是新推出的无线传感器协议之一，能够远距离传送消息并消耗低能量，这些特性使其适用于低成本设备。虽然 LoRa 是作为固定设备的一项技术引入的，但它也可以用于低速移动设备。在本文中，我们通过提高移动终端设备的连接可靠性，同时将能耗保持在低水平，引入了对自适应数据速率 (ADR) 机制的增强，以支持移动 LoRa。首先，我们为网络服务器端提出了线性回归-ADR (LR-ADR) 机制，以平滑每个网关的信噪比 (SNR) 估计并预测下一次传输的 SNR。其次，我们提出了线性回归+ADR（LR+ADR）机制，这是一种自适应方法，用于终端设备端更快地重新获得与网络服务器的连接。我们进行了仿真建模以评估我们实施的性能，同时将我们的结果与四种替代解决方案 ADR、ADR+、EMA-ADR、G-ADR 进行比较。结果证明我们的第一种方法 (LR-ADR) 的性能优于最佳竞争对手，而我们的第二种方法 (LR+ADR) 在数据包交付率 (PDR) 方面带来了额外的改进，同时它们保留了每个数据包的能耗以低水平交付 (ECPD)。特别是，在模拟真实世界条件的场景中，LR+ADR 的 PDR 与原始 ADR 相比提高了 520%，与最佳竞争对手 (G-ADR) 相比提高了 38%。此外，与原始 ADR 相比，它可将 ECPD 降低多达 74%，同时与最佳竞争对手 (G-ADR) 保持在同一水平。