Industry 4.0, Industrial Internet of Things, Cyber-Physical Production Systems and Smart Factories are closely related emerging concepts expected to drive significant improvements in industrial production systems, with gains in efficiency, cost and customer satisfaction. These concepts are intimately associated with highly distributed and cooperative architectures that rely, naturally, on the network infrastructure. However, traditional industrial communication technologies hardly provide the required level of integration, flexibility and performance. Seeking a solution to this mismatch, we assess two technologies that appeared recently in the industrial realm, namely IEEE 802.1 Time-Sensitive Networking (TSN) and Software-Defined Networking (SDN). TSN and SDN are fundamentally different, thus having distinct strengths and weaknesses. This paper reviews their fundamental operation principles, evaluating them qualitatively against the specific requirements posed by Industry 4.0.

Floor-cleaning robots are becoming increasingly more sophisticated over time and with the addition of digital cameras supported by a robust vision system they become more autonomous, both in terms of their navigation skills but also in their capabilities of analyzing the surrounding environment. This document proposes a vision system based on the YOLOv5 framework for detecting dirty spots on the floor. The purpose of such a vision system is to save energy and resources, since the cleaning system of the robot will be activated only when a dirty spot is detected and the quantity of resources will vary according to the dirty area. In this context, false positives are highly undesirable. On the other hand, false negatives will lead to a poor cleaning performance of the robot. For this reason, a synthetic data generator found in the literature was improved and adapted for this work to tackle the lack of real data in this area. This synthetic data generator allows for large datasets with numerous samples of floors and dirty spots. A novel approach in selecting floor images for the training dataset is proposed. In this approach, the floor is segmented from other objects in the image such that dirty spots are only generated on the floor and do not overlap those objects. This helps the models to distinguish between dirty spots and objects in the image, which reduces the number of false positives. Furthermore, a relevant dataset of the Automation and Control Institute (ACIN) was found to be partially labelled. Consequently, this dataset was annotated from scratch, tripling the number of labelled images and correcting some poor annotations from the original labels. Finally, this document shows the process of generating synthetic data which is used for training YOLOv5 models. These models were tested on a real dataset (ACIN) and the best model attained a mean average precision (mAP) of 0.874 for detecting solid dirt. These results further prove that our proposal is able to use synthetic data for the training step and effectively detect dirt on real data. According to our knowledge, there are no previous works reporting the use of YOLOv5 models in this application.

This paper investigates the usage of machine learning algorithms, applied to the task of position estimation in visible light positioning systems. Traditional approaches relying in trilateration usually resort to the application of the least squares method to find the position estimate. The least squares method is very prone to outlier information present in the data set, which reduces the estimation accuracy. This paper presents a strategy based on clustering and outlier removal able to improve the estimation accuracy. Clustering is based on DBSCAN, an algorithm used to find structure in unstructured data. The tuning parameters for DBSCAN are optimized following a linear regression supervised learning step, where a set of training examples with known real position is used. Simulation results show a 35% gain improvement in accuracy achieved with a moderate complexity increase. The minimum estimation error for the case scenario under study was 0.2 mm, with an r.m.s. error of 35 mm.

This work demonstrates experimentally the feasibility of using up to 32‐PAM in rolling shutter based optical camera communication (OCC). Two different cameras have been used for OCC performance comparison, namely, a low‐cost Pi Camera V2.1 and a higher‐cost Alvium 1800 U‐500m. Based on the limitations evidenced by this analysis, a decoding algorithm has been proposed to recover the transmitted signal from the acquired images. The experimental evaluation has been carried out for distances up to 2 m, using 8‐PAM, 16‐PAM and 32‐PAM, analysing the received samples histograms and the symbol error rate (SER) for both cameras. Based on the collected experimental results, it is concluded that 32‐PAM is possible with both cameras, with a maximum SER of 0.16% for the Alvium camera and 5.05% for the Pi Camera. Finally, it is observed that both the signal and noise amplitudes remain constant for the tested distances, resulting in a SER that does not notably depend on the distance between the transmitter and receiver.