Analysis of Research Trends in Fractional Controller Using Latent Dirichlet Allocation

Complex industrial processes demand high performance control characteristics. Fractional order controller is more robust than the classical controllers generally used in industry. This paper presents an analysis of research trends in this ever-growing domain. Firstly, a corpus of 200 research articles published during 2014-19 was collected from selected journals. Latent Dirichlet Allocation methodology was applied on this literature data set for topic modeling. Optimal number of topics, average topic weights and topic prevalence per year were obtained. Distributions of topic proportions were generated for the selected journals and with respect to the author nationalities. In addition, continent and country-wise distribution of documents were also depicted. Primary results indicate that the topics of fractional order controller design and stability of fractional order system have been researched most in past years, but are on a declining trend. Fractional sliding control system and applications of fractional order systems are upward trending topics. Other topics like optimization of fractional order systems need more attention. In addition, journals publishing in the upward trending topics were identified. Journals publishing in a wide range of topics were also determined. It was found that almost 75% of all fractional controller publications during 2014-19 have Asian first authors, followed by European and African first authors. The majority of first authors belong to China (31%), India (19%), Iran(13%)and Turkey (8%). With regards to topic diversity, the Chinese and Indian authors have worked on the widest range of fractional controller topics, followed by the authors from Algeria, Iran, Portugal, Spain and Turkey. Researchers from other nations have focused on fewer topics.