Evaluation method for application of internet of things for aquaculture
X Yuan,Q Kong,Q Li,L Li,D Li
DOI: https://doi.org/10.3969/j.issn.1002-6819.2015.04.036
2015-01-01
Abstract:Internet of things for Aquaculture is an integrated modern system based on computing and communications technology like smart sensor technology, reliable telecommunication, intelligent information processing, which can collect data and images, transmit and process data intelligently, forecast future trend and early-warning for decision support. First of all, it is a key issue to establish scientific and rational index system and evaluation method for internet of things for aquaculture to guarantee its effectiveness. With the rapid development of information technology in China, the internet of things for aquaculture has been promoted and applied in Jiangsu, Shandong, Hunan, Hubei, Zhejiang and Guangdong. However, the internet of things for aquaculture in China is at an early stage and there are some problems, which cause negative impact on the promotion and application of the system, for examples, redundant functions, high cost, unstable performances, and so on. Secondly, in order to assess internet of things for aquaculture system, index system to assess internet of things for aquaculture was built in this paper by indicator optimization model to solve randomicity, redundancy, cross-connection and overlap caused subjective selection. Three steps composed of the selection process: 1) first round selection, three categories indices of function indicator, performance indicator and effectiveness indicator, targeted to 40 indicators were selected; 2) second round selection, 40 indicators representing perception layer, transmission layer and application layer, were optimized to 26 by method proposed by Dale and Beyeler, in which the standard conformity degree of each indicator was checked one by one and indicators need to meet at least 5 standards, otherwise they will be eliminated; 3) indicator screening model, by which 26 indicators were reduced to 14, with only 35% of total indicators representing 88.45% of total information, capturing the requirements of completeness and simplicity. Thirdly, fuzzy comprehensive evaluation approach, which comprise of three levels fuzzy comprehensive evaluation, was established to assess the application of internet of things for aquaculture. Most of the 14 indicators got in the first phase are qualitative factors and difficult to be quantified. That is why fuzzy comprehensive evaluation approach was used in this paper. Meanwhile, it is better to adopt multi-level factors as weights are not easy to be assigned reasonably when factors are too many, which leads unreasonable and wrong results. Multi-level fuzzy comprehensive method works well to solve this problem and can be applied to assess function characteristics, performance and effectiveness of the system of internet of things for aquaculture. Multi-level fuzzy comprehensive was built by steps of establishing comment set, membership and got quantitative scoreA, which can fall into five levels of excellence (A≥82.5), good (82.5>A≥67.5), satisfactory (67.5>A≥52.5), barely adequate (52.5>A≥35) and fail (A<35). Finally, case studies were carried out in Yixing of crab cultivation, Jiangsu province and Zhanjiang of white shrimp cultivation, Guangdong province, in which the index system and multi-level fuzzy comprehensive method were tested based on the internet of things for aquaculture system developed independently by China Agricultural University (CAU). Results showed that the score in Yixing was 87.371 indicating application of internet of things for aquaculture system in Yixing is excellent, meanwhile, the score in Zhanjiang is 74.921, which is good. Both of the two results are consistent with the actual situation of two bases, showing that index system and multi-level fuzzy comprehensive method proposed in this paper are feasible and reasonable for evaluation of application of internet of things for aquaculture and can guide its construction and improvement.