International audience

In today's world, there is a growing need to analyze data stored in a Data Lake, which is a collection of large, heterogeneous databases. Our work is part of a medical application that aims to help healthcare professionals analyze complex data for decision-making. We propose mechanisms that promote data accessibility. The data are stored in a Data Warehouse (DW) that is periodically built from a data lake. Depending on the needs of the decision-maker, data are extracted from the DW and transferred to a Data Mart (DM) for querying. In this paper, we present a schema recommendation system based on the principle of collaborative filtering. This system can predict the DM schemas that were developed in the past that best match the data need expressed by a decision-maker. It does this by comparing the attributes present in the schemas with the attributes deduced from the need to propose a list of predictions for the most suitable schemas. The technique used is simple, while allowing us to solve the problem of periodic updates to the source data. An experiment was conducted for a medical application.

Massive databases or Big Data are generally managed by NoSQL systems. Most of these systems are schemaless, meaning that the schema is not predefined; it is provided as the database (DB) is fed. This property increases data flexibility. However, the absence of a schema is a major obstacle for expressing complex queries and for the semantic analysis of data. In our previous work [1], we proposed a process for extracting the logical schema of a document-oriented NoSQL DB; this schema allows to describe the structure of the data stored in the DB and thus facilitates query writing. But to understand the meaning of the data and its structure, a conceptual schema is the most appropriate tool, since it ignores the technical aspects linked to implementation. In this paper, we propose a process for transforming the logical schema of a DB into a UML-type conceptual schema. It shows the object classes and semantic links (association, composition and inheritance) contained in the DB. Based on Model Driven Architecture (MDA), our process relies on metamodels and a set of transformation rules to automatically produce the conceptual schema from the logical schema. We experimented our process using a medical application.

When a robot performs a motion control task in a real dynamic environment, it has to account not only for maintaining the track where it is, and how it expects to rise, but also to ensure that dynamic uncertainties and task specific limitations are properly expected. In this paper, a second order sliding mode controller has been implemented into a robotic system for a trajectory tracking task, in the case of ideal functioning as well as for real systems submitted to parameters uncertainties. Then, an adaptive extension of the second order sliding mode control has been established to tackle the presence of physical environmental disturbances and mainly the parametric uncertainties problems. An experimental implementation of the adaptive second order SMC performed on 4 DOF Lynx robot illustrates the effectiveness of the architecture control.

International audience

Volume, Variety and Velocity are the three dimensions that have definitely changed the tools we need to store and process Big Data effectively, giving rise to NoSQL systems for faster data access, better scalability and higher flexibility. While NoSQL systems have proven their efficiency to handle Big Data, it is still an unsolved problem how the automatic storage of Big Data in NoSQL systems could be done. One solution for addressing this problem is to model Big Data, and then define mapping rules towards the physical level. This paper proposes an automatic MDA-based approach that translates conceptual models expressed using the Unified Modeling Language (UML) into NoSQL physical models. Our approach rely on an intermediate logical model compatible with column, document and graph oriented systems which allows to choose the system type that suits the best with business rules and technical constraints.