农业生态环境监测中无线传感节点信号有效传输距离的确定

农业生态环境的物理形态和结构复杂多样，对WSN（wireless sensor networks）的无线信号传输造成不同衰减影响。为确保无线传感器网络在农业环境中经济、合理、高效部署，有必要明确典型农业环境中无线传感节点间的有效传输距离。该文基于Shadowing信号衰减模型，利用当前通用的CC2530和CC2591无线通信模块，分别选定4种不同农业环境（湖泊、草地、农田、树林）开展单跳组网试验，通过设定不同距离测试传感器节点的接收信号强度指标（received signal strength indication，RSSI），分析不同环境中RSSI与传输距离间的变化特征。试验结果表明，所有测试环境获得的RSSI值与有效距离遵从Shadowing模型，其拟合度在0.9232～0.9846之间。通过对实测数据建立拟合模型，以接收节点的灵敏度为临界值，计算出湖泊、草地、农田、树林4种环境的理论传输距离分别为663.3，419.3，208.0和79.5 m，而实测有效传输距离与理论值之间的相对误差在22%～34%之间。从误差分布看，复杂环境的实测值更接近理论值，而特殊结构的复杂环境似对实际信号传输有增强作用。该文的研究方法和模型估算获得的信号衰减系数可为实际环境监测组网提供有益参考。

Determination of effective transmission distances of wireless sensor network nodes for agro-ecological environment monitoring

Abstract: Wireless sensor networks (WSN) have been widely adopted for monitoring of agro-ecological environment, as they offer a number of advantages over traditional field observation methods. Signal transmission distances and qualities achieved by wireless sensors are highly related to the types of external environments. Attenuation of radio signals varies drastically for wireless sensor networks in different agro-ecological environments with diverse physical forms and structures. To achieve the economic, rational, and efficient goal for WSN deployment, it is essential to identify the effective transmission distance between wireless sensors in typical agro-ecological environments. This paper employed a long attenuation model, known as a Shadowing model, to examine the effect of distance on signal propagation loss with given transmitting power by measuring signal strength at the receiving node in one-hop networking experiments. The network was constructed using 12 nodes with commonly adopted CC2530 and CC2591 as wireless communication modules (both working at 2.4GHz ISM frequency band) and four different landscape settings as typical analysis environments (i.e. lake, grassland, low shrubs, and woodland). To improve the signal sending and receiving capacity, wireless sensor nodes under experiment were equipped with a 5 dB short-stick antenna. The initial test was conducted to determine the minimum sensitivity of sensor nodes to be -97 dBm. By setting a series of distances between data sending and receiving sensor nodes, the corresponding received signal strength indication (RSSI) was recorded. A Matlab-based nonlinear regression model was built with the recorded RSSI data to analyze the relationship between RSSI and transmission distance for each of the four agro-ecological environments. The resulting coefficients of determination for the regression models indicated a strong relationship between RSSI and transmission distance, as they complied with the Shadowing model with a degree of fitting between 0.9232-0.9556. According to the fitted curves in the regression analyses, a transmission path loss index was calculated to represent such interferences as attenuation, reflection, and multi-path phenomenon on wireless signals due to a given environmental morphology and structure. With the signal sending power being kept constant, it was found that lakes had the lowest transmission path loss index value (2.1), which was followed by grassland (2.4), low shrubs (2.6), and finally woodland (3.1). Based on the fitted regression models and adopting the minimum sensitivity of sensors as the threshold, the calculated theoretical maximum transmission distance for the deployment of the given sensor nodes was 663.3m, 419.3m, 155.2m, and 79.5m for lake, grassland, low shrubs, and woodland, respectively. Effective transmission distances were also computed from the theoretical ones by a 25% deduction, resulting in 495m, 330m, 150m and 65m for the above four ecological settings, respectively. The experimental procedure and the transmission path loss index estimated by the fitted regression models in this paper can provide useful reference for practical environmental monitoring network construction and sensor node deployment when facing diverse environmental morphologies. Potential further work will include investigating effects of node height variation on transmission distances in different agro-ecological environments and experimenting and comparing the results of this study with sensor nodes equipped with other wireless modules.