Since its beginning, around the 50s decade, until present days, the area of unmanned underwater vehicles (UUV) has considerably grown through time; those have been used for many tasks and applications, from bomb searching and recovery to sea exploration. Initially, these robots were used mainly for military and scientific purposes. However, nowadays, they are very much extended into civils, and it is not hard to find them being used for recreation. In this context, the present research is an effort to make a walkthrough of evolution in this area, showing a diversity of structure designs, used materials, sensor and instrumentation technologies, kinds and the number of actuators employed, navigation control techniques, and what is new in development trends. The paper gives a clear starting point for those who are initializing into this research area; also, it brings some helpful knowledge for those who already have experience.

Technological advances in the field of Remote Sensing generate large volumes of geospatial data. Current geographic information systems (GIS) doesn't support the massive processing of satellite imagery, two examples of this kind of software are: (i) the Brazilian Spring project, GIS and remote sensing image processing system (ii) QGIS, a free and open source GIS. To achieve massive processing, we have HIPI framework, it provides a solution for how to store a large collection of images and works on the Hadoop Distributed File System. Currently, HIPI only supports specific image formats, such as, JPEG, PNG and PPM. In this article is presented a new approach to distributed processing of considerable amounts of satellite images. We make an extension of HIPI to support satellite images format, TIFF, this fact helps to preserve the information, process and analyze satellite images massively to have results faster than the traditional way.

Actualmente la Covid-19 está causando grandes pérdidas a nivel mundial, es por ello que diferentes trabajos permiten predecir o pronosticar el comportamiento de cantidad de infectados utilizando técnicas de forecasting, dentro del campo de Inteligencia Artificial se están permitiendo tomar medidas de control en los diferentes países. En este trabajo se propuso un modelo de aprendizaje profundo para pronosticar los casos diarios en las regiones de Madre de Dios. Los datos utilizados pertenecen al conjunto de datos abiertos Covid-19 del Ministerio de Salud de Perú (MINSA). El conjunto de datos incluye los períodos de inicio de marzo de 2020 a fines de diciembre de 2021. Se utilizó un LSTM utilizando variables de Fecha, Departamento, Provincia, Distrito, Casos, IP. ID y con tamaño de ventana de 5 días, se obtuvo una precisión de 94,67% con los datos de entrenamiento y un 92,31%.