The sector of agriculture facing numerous challenges for the proper utilization of its natural resources. For that reason, and to the growing risk of changing weather conditions, we must monitor the soil conditions and meteorological data locally in order to accelerate the adoption of appropriate decisions that help the culture. In the era of the Internet of Things (IoT), a solution is to deploy a Wireless Sensor Network (WSN) as a low-cost remote monitoring and management system for these kinds of features. But WSN is suffering from the motes’ limited energy supplies, which decrease the total network’s lifetime. Each mote collects periodically the tracked feature and transmitting the data to the edge Gateway (GW) for further study. This method of transmitting massive volumes of data allows the sensor node to use high energy and substantial usage of bandwidth on the network. In this research, Data Traffic Management based on Compression and Minimum Description Length (MDL) Techniques is proposed which works at the level of sensor nodes (i.e., Things level) and at the edge GW level. In the first level, a lightweight lossless compression algorithm based on Differential Encoding and Huffman techniques which is particularly beneficial for IoT nodes, that monitoring the features of the environment, especially those with limited computing and memory resources. Instead of trying to formulate innovative ad hoc algorithms, we demonstrate that, provided general awareness of the features to be monitored, classical Huffman coding can be used effectively to describe the same features that measure at various time periods and locations. In the second level, the principle of MDL with hierarchical clustering was utilized for the purpose of clustering the sets of data coming from the first level. The strategy used to minimize data sets transmitted at this level is fairly simple. Any pair of data sets that can be compressed according to the MDL principle is combined into one cluster. As a result of this strategy, the number of data sets is gradually decreasing and the process of merging similar sets into a single cluster is stopped if no more pairs of sets can be compressed. Results utilizing temperature measurements indicate that it outperforms common methods developed especially for WSNs in reducing the amount of data transmitted and saving energy, even though the suggested system does not reach the theoretical maximum.

农业部门在合理利用其自然资源方面面临众多挑战。因此，为了适应不断变化的天气状况，我们必须在当地监测土壤状况和气象数据，以加快采用有助于文化发展的适当决策。在物联网（IoT）时代，一种解决方案是将无线传感器网络（WSN）部署为这些功能的低成本远程监视和管理系统。但是WSN受到尘埃有限的能源供应的困扰，这减少了整个网络的寿命。每个节点定期收集跟踪的功能，并将数据传输到边缘网关（GW）进行进一步研究。这种传输大量数据的方法允许传感器节点使用高能量并大量使用网络上的带宽。在这项研究中，提出了基于压缩和最小描述长度（MDL）技术的数据流量管理，该技术可在传感器节点级别（即事物级别）和边缘GW级别工作。在第一级中，基于差分编码和霍夫曼技术的轻量级无损压缩算法对物联网节点特别有益，它可以监视环境的特征，尤其是那些计算和内存资源有限的特征。我们没有尝试制定创新的即席算法，而是证明了，只要能够全面了解要监视的功能，传统的霍夫曼编码可以有效地描述在不同时间段和位置测量的相同特征。在第二级中，利用具有分层聚类的MDL原理来聚类来自第一级的数据集。用于最小化在此级别传输的数据集的策略非常简单。可以根据MDL原理压缩的任何一对数据集都组合到一个群集中。这种策略的结果是，如果无法压缩更多对数据集，则数据集的数量将逐渐减少，并且停止将相似数据集合并为单个群集的过程。利用温度测量的结果表明，在减少传输的数据量和节省能源方面，它的性能优于为WSN开发的常规方法。