Mateus Mendes

Artificial neural network modelling of the antioxidant activity and phenolic compounds of bananas submitted to different drying treatments

Bananas (cv. Musa nana and Musa cavendishii) fresh and dried by hot air at 50 and 70°C and lyophilisation were analysed for phenolic contents and antioxidant activity. All samples were subject to six extractions (three with methanol followed by three with acetone/water solution). The experimental data served to train a neural network adequate to describe the experimental observations for both output variables studied: total phenols and antioxidant activity. The results show that both bananas are similar and air drying decreased total phenols and antioxidant activity for both temperatures, whereas lyophilisation decreased the phenolic content in a lesser extent. Neural network experiments showed that antioxidant activity and phenolic compounds can be predicted accurately from the input variables: banana variety, dryness state and type and order of extract. Drying state and extract order were found to have larger impact in the values of antioxidant activity and phenolic compounds.

Convective Drying of Apples: Kinetic Study, Evaluation of Mass Transfer Properties and Data Analysis using Artificial Neural Networks

Abstract In the present work, the effect of drying was evaluated on some chemical and physical properties of apples, and the functions were modelled using feed-forward artificial neural networks. The drying kinetics and the mass transfer properties were also studied. The results indicated that acidity and sugars were significantly reduced by drying. Regarding colour lightness decreases, whereas redness and yellowness increased. As for texture, the dried samples were softer and less cohesive as compared to the fresh ones. Mass diffusivity increased with temperature, from 4.4×10−10 m2/s at 30°C to 1.4×10−9 m2/s at 60°C, and so did the mass transfer coefficient, increasing from 3.7×10−10 m/s at 30°C to 7.4×10−9 m/s at 60°C. As to the activation energy, it was found to be 34 kJ/mol. Neural network modelling showed that all properties can be correctly predicted by feed-forward neural networks. The analysis of the networks’ behaviours input layer weight values also shows which properties are more affected by dehydration or more dependent on variety.

Artificial neural network modelling of the chemical composition of carrots submitted to different pre-drying treatments

The effect of various pre-drying treatments on the quality of dried carrots was evaluated by assessing the values of moisture, ash, protein, fibre, sugars and colour. The pre-drying treatments under investigation were dipping, either in ascorbic acid or sodium metabisulphite at different concentrations and pre-treatment times, as well as blanching. The experimental data was analysed using neural networks, so that relevant patterns in the data were found and conclusions drawn about each variable. The results showed that the type of pre-drying treatment (chemical or physical) had variable impact on the nutritional composition of the dried carrots but not on the colour parameters, which were found to be mostly unaffected by the pre-treatment procedure. Pre-treatment with chemical agents such as ascorbic acid or metabisulphite seem to have the least impact on the parameters studied. The results of the analysis by artificial neural networks confirmed these findings.

Robot navigation based on view sequences stored in a sparse distributed memory

Robot navigation is a large area of research, where many different approaches have already been tried, including navigation based on visual memories. The Sparse Distributed Memory (SDM) is a kind of associative memory based on the properties of high-dimensional binary spaces. It exhibits characteristics, such as tolerance to noise and incomplete data, ability to work with sequences and the possibility of one-shot learning. Those characteristics make it appealing to use for robot navigation. The approach followed here was to navigate a robot using sequences of visual memories stored into a SDM. The robot makes intelligent decisions, such as selecting only relevant images to store, adjusting memory parameters to the level of noise and inferring new paths from the learnt trajectories. The method of encoding the information may influence the tolerance of the SDM to noise and saturation. This paper reports novel results of the limits of the model under different typical navigation problems. The SDM showed to be very robust to illumination and scenario changes, occlusion and saturation. An algorithm to build a topological map of the environment based on the visual memories is also described.

Product Traceability in Ceramic Industry 4.0: A Design Approach and Cloud-Based MES Prototype

We propose a customer-focused approach to design product traceability for Industry 4.0. Our design-science research includes a review of traceability technologies and participative enterprise modeling in the ceramic industry. We find benefits in combining Business Process Modeling Notation and Goal-oriented Requirements Language representations to (1) promote reflection by experts with different backgrounds, (2) reach consensus with a solution that addresses the goals of multiple stakeholders, and (3) ensure that customers’ needs are a priority in traceability design. The resulting model combines technologies in different stages of the product lifecycle and is implemented in a cloud-based MES (Manufacturing Execution System) prototype. Depending on each stage and strategic intention, the identification code can be embedded in the product, transport, or package. Our contribution can assist managers in the creation of cloud-based MES to support traceability integration at (1) technological, (2) vertical, and (3) horizontal levels that are required in the fourth industrial revolution.

Vision-Based Collision Avoidance for Service Robot

ASSIS is a service robot which uses a camera, sonars and infra-red sensors for navigation. It uses images stored into a Sparse Distributed Memory, implemented in parallel in a Graphics Processing Unit, as a method for robot localization and navigation. It is controlled from a web-based interface. Algorithms for following previously learnt paths using visual and odometric information are described. A stack-based method for avoiding random obstacles, using visual information, is proposed. The results show the algorithms are adequate for indoors robot localization and navigation.

Experiments with a Sparse Distributed Memory for Text Classification

The Sparse Distributed Memory (SDM) has been studied for decades as a theoretical model of an associative memory in many aspects similar to the human brain. It has been tested for different purposes. The present work describes its use as a quick text classifier, based on pattern similarity only. The results found with different datasets were superior to the performance of the dumb classifier or purely random choice, even without text preprocessing. Experiments were performed with a popular Reuters newsgroups dataset and also for real time web ad serving.

Topological Mapping Using Vision and a Sparse Distributed Memory

Navigation based on visual memories is very common among humans. However, planning long trips requires a more sophisticated representation of the environment, such as a topological map, where connections between paths are easily noted. The present approach is a system that learns paths by storing sequences of images and image information in a sparse distributed memory (SDM). Connections between paths are detected by exploring similarities in the images, using the same SDM, and a topological representation of the paths is created. The robot is then able to plan paths and switch from one path to another at the connection points. The system was tested under reconstitutions of country and urban environments, and it was able to successfully map, plan paths and navigate autonomously.

Encoding Data to Use with a Sparse Distributed Memory

A Sparse Distributed Memory (SDM) is a kind of associative memory suitable to work with high-dimensional vectors of random data. This memory model exhibits the characteristics of a large boolean space, which are to a great extent those of the human long-term memory. Hence, this model is attractive for Robotics and Artificial Intelligence, since it can possibly grant artificial machines those same characteristics. However, the original SDM model is appropriate to work with random data. Sensorial data is not always random: most of the times it is based on the Natural Binary Code and tends to cluster around some specific points. This means that the SDM performs poorer than expected. As part of an ongoing project, in which the goal is to navigate a robot using a SDM to store and retrieve sequences of images and associated path information, different methods of encoding the data were tested. Some methods perform better than others, and one method is presented that can offer the best performance and still maintain the characteristics of the original model.