This work presents a framework for coffee maturity classification from multispectral image data based on Convolutional Neural Networks (CNNs). The system leverages the use of multispectral image acquisition systems that generate large amounts of data, by taking advantage of the ability of CNNs to extract meaningful patterns from very high-dimensional data. We validated the use of five different popular CNN architectures on the classification of cherry coffee fruits according to their ripening stage. The different models were trained on a training dataset balanced in different ways, which resulted in a top accuracy higher than 98% when applyed to the classification of 600 coffee fruits in 5 different stages of ripening. This work has the potential of providing the farmer with a high-quality, optimized, accurate and viable method for classifying coffee fruits. In order to foster future research in this area, the data used in this work, which was acquired with a custom-developed multispectral image acquisition system, has been released.

Reliable and robust control of power converters is a key issue in the performance of numerous technological devices. In this paper we show a design technique for the control of a DC-DC buck converter with a switching technique that guarantees both good performance and global stability. We show that making use of the contraction theorem in the Jordan canonical form of the buck converter, it is possible to find a switching surface that guarantees stability but it is incapable of rejecting load perturbations. To overcome this, we expand the system to include the dynamics of the voltage error and we demonstrate that the same design procedure is not only able to stabilize the system to the desired operation point but also to reject load, input voltage, and reference voltage perturbations.

Multispectral imaging systems using narrow-band power LEDs have become a feasible solution for a wide range of applications. Their fast response and simple implementation compensates for a low spectral resolution. In this paper, we are filling a gap identified in the literature on multispectral systems by presenting a detailed design of a light controller, as well as a characterization of the whole system by means of a correlation and a repeatability analysis. We use a reference color target to assess accuracy and precision by comparing our system with high-performance hyperspectral equipment used in fruit quality control. The system captures multispectral images with 15 different wavelengths (λ∈[410 950] nm). The main finding is a Pearson correlation index above 80% (ρ>0.8) with a variance in the measure within 10% (σ2<0.1). We also examine the characteristics of the instrument at a maximum speed of 8 multispectral images per second, although this entails sacrificing precision (σ2<0.3).

The use of optimal control theory for motion planning is a challenging task. Cell mapping offers a way to formulate combinatorial optimization problems, allowing the inclusion of complex cost functions as well as multi-objective optimization problems. This paper presents a suboptimal solution for a trajectory planning problem in a workspace with obstacles, for a differential drive mobile robot. This method relies on the use of any linearization technique that allows the regularization of the combinatorial optimization problem. We explore some classical problems in optimal control, i.e., distance, control effort, and navigation time), as well as the multi-objective optimization problem (MOP). We also performed a comparison with two classical path planning algorithms, namely $A^{\ast }$ and $RR{T}^{\ast }$, to validate the proposed method when the multi-objective optimization problem includes distance in the cost function, achieving a compromise of less than $2\%$ for the worst-case scenario for our case study.

High voltage gain power converters are very important in photovoltaic applications mainly due to the low output voltage of photovoltaic arrays. This kind of power converters includes three or more semiconductor devices and four or more energy storage elements, making the dynamical analysis of the controlled system more difficult. In this paper, the boost-flyback power converter is controlled by peak-current mode with compensation ramp. The closed-loop analysis is performed to guarantee operation conditions such that a period-1 orbit is attained. The converter is considered as a piecewise linear system, and the closed-loop stability is determined by using the monodromy matrix, obtained by the composition of the saltation matrixes with the solutions of the dynamical equations in the linear intervals. The largest eigenvalue of the monodromy matrix gives the stability of the period-1 orbit, and a deep analysis using bifurcation diagrams let us reach a conclusion about the loss of the stability, which is experimentally verified. To avoid overcompensation effects, the minimum value required by the compensation ramp is obtained, and the minimum and maximum values of the load resistance are found too. The system has a good transient response under disturbances in the load and in the input voltage.