Abstract
The article provides a comparison of Convolutional Neural Network (CNN) and Reinforcement Learning (RL) applied to the field of autonomous driving within the CARLA (CAr Learning to Act) simulator for training and evaluation. The analysis of results revealed CNNs better overall performance, as it demonstrated a more refined driving experience, shorter training durations, and a more straightforward learning curve and optimization process. However, it required data labelling. In contrast, RL relayed on an exhaustive (unsupervised) exploration of different models, ultimately selecting the model at timestep 600,000, which had the highest mean reward. Nevertheless, RL’s approach revealed its susceptibility to excessive oscillations and inconsistencies, necessitating additional optimization and tuning of hyperparameters and reward functions. This conclusion is further substantiated by a range of used performance metrics (objective and subjective), designed to assess the performance of each approach. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

Abstract
This study focuses on the analysis of emulsion pictures to understand important parameters. While droplet size is a key parameter in emulsion science, manual procedures have been the traditional approach for its determination. Here we introduced the application of YOLOv7, a recently launched deep-learning model, for classifying emulsion droplets. A comparison was made between the two methods for calculating droplet size distribution. One of the methods, combined with YOLOv7, achieved 97.26% accuracy. These results highlight the potential of sophisticated image-processing techniques, particularly deep learning, in chemistry-related topics. The study anticipates further exploration of deep learning tools in other chemistry-related fields, emphasizing their potential for achieving satisfactory performance.

Abstract
Monitoring and controlling the energy consumption of electrical appliances brings significant benefits to both consumers and the energy utility. This work presents a system for monitoring and controlling energy consumption by residence loads connected to smart plugs. The user will have a tool to view consumption information and remotely turn loads on and off, as well as control the power level at which certain appliances will operate. In addition, it is intended to give the system the ability to make decisions regarding the operation of electrical devices based on the electrical energy available. This decision-making can occur either through priorities established by the user or, possibly, through Machine Learning applied to the system, based on the consumption pattern. Solutions like these can even be applied in situations where the user produces his own energy and would like to use the surplus produced to meet certain loads.

Abstract

Abstract
Data acquisition by electromyography, as well as the muscle stimulation, has become more accessible with the new developments in the wearable technology and medicine. In fact, for treatments, games or sports, it is possible to find examples of the use of muscle signals to analyse specific aspects related, e.g., to disease, injuries or movement impulses. However, these systems are usually expensive, does not integrate data acquisition with the muscle stimulation and does not exhibit an adaptive control behaviour that consider the pathology and the patient response. This paper presents a wearable system that integrates the signal acquisition and the electrostimulation using dry thin-film titanium-based electrodes. The acquired data is transmitted to a mobile application running on a smartphone by using Bluetooth Low Energy (BLE) technology, where it is analysed by employing artificial intelligence algorithms to provide customised treatments for each patient profile and type of pathology, and taking into consideration the feedback of the acquired electromyography signal. The acquired patient's data is also stored in a secure cloud database to support the physician to analyse and follow-up the clinical results from the rehabilitation process.

Abstract
Abstract Background: The association of distributed generators, energy storage systems and controllable loads close to the energy consumers gave place to a small-scale electrical network called microgrid. The stochastic behavior of renewable energy sources, as well as the demand variation, can lead in some cases to problems related to the reliability of the microgrid system. On the other hand, the market price of electricity from mainly non-renewable sources becomes a concern for a simple consumer due to its high costs.Method: In this work, an energy management system was developed based on an innovative optimization method, combining linear programming, based on the simplex method, with particle swarm optimisation algorithm. Two scenarios have been proposed to characterise the relation price versus gas emissions for optimal energy management. The objective of this study is to nd the optimal setpoints of generators in a smart city supplied by a microgrid in order to ensure consumer comfort, minimising the emission of greenhouse gases and ensure an appropriate operating price for all smart city consumers. Results: The simulation results have demonstrated the reliability of the optimisation approach on the energy management system in the optimal scheduling of the microgrid generators power ows, having achieved a better energy price compared to a previous study with the same data. Conclusion: The energy management system based on the proposedoptimisation approach gave an inverse correlation between economic and environmental aspects, in fact, a multi-objective optimisation approach is performed as a continuation of the work proposed in this paper.