The analysis and follow up of asphalt infrastructure using image processing techniques has received increased attention recently. However, the vast majority of developments have focused only on determining the presence or absence of road damages, forgoing other more pressing concerns. Nonetheless, in order to be useful to road managers and governmental agencies, the information gathered during an inspection procedure must provide actionable insights that go beyond punctual and isolated measurements: the characteristics, type, and extent of the road damages must be effectively and automatically extracted and digitally stored, preferably using inexpensive mobile equipment. In recent years, computer vision acquisition systems have emerged as a promising solution for road damage automated inspection systems when integrated into georeferenced mobile computing devices such as smartphones. However, the artificial intelligence algorithms that power these computer vision acquisition systems have been rather limited owing to the scarcity of large and homogenized road damage datasets. In this work, we aim to contribute in bridging this gap using two strategies. First, we introduce a new and very large asphalt dataset, which incorporates a set of damages not present in previous studies, making it more robust and representative of certain damages such as potholes. This dataset is composed of 18,345 road damage images captured by a mobile phone mounted on a car, with 45,435 instances of road surface damages (linear, lateral, and alligator cracks; potholes; and various types of painting blurs). In order to generate this dataset, we obtained images from several public datasets and augmented it with crowdsourced images, which where manually annotated for further processing. The images were captured under a variety of weather and illumination conditions and a quality-aware data augmentation strategy was employed to filter out samples of poor quality, which helped in improving the performance metrics over the baseline. Second, we trained different object detection models amenable for mobile implementation with an acceptable performance for many applications. We performed an ablation study to assess the effectiveness of the quality-aware data augmentation strategy and compared our results with other recent works, achieving better accuracies (mAP) for all classes and lower inference times (3× faster).

In this paper, a real-time implementation of a sliding-mode control (SMC) in a hardware-in-loop architecture is presented for a robot with two degrees of freedom (2DOF). It is based on a discrete-time recurrent neural identification method, as well as the high performance obtained from the advantages of this architecture. The identification process uses a discrete-time recurrent high-order neural network (RHONN) trained with a modified extended Kalman filter (EKF) algorithm. This is a method for calculating the covariance matrices in the EKF algorithm, using a dynamic model with the associated and measurement noises, and it increases the performance of the proposed methodology. On the other hand, the decentralized discrete-time SMC technique is used to minimize the motion error. A Virtex 7 field programmable gate array (FPGA) is configured based on a hardware-in-loop real-time implementation to validate the proposed controller. A series of several experiments demonstrates the robustness of the algorithm, as well as its immunity to noise and the inherent robustness to external perturbation, as are typically found in the input reference signals of a 2DOF manipulator robot.

In this paper the hardware implementation of a NARX neural network algorithm using a Field Programmable Gate Array (FPGA) is presented. A NARX network is a Recurrent Neural Network (RNN) suitable for modeling nonlinear systems with promising results for the modeling of the inverse characteristics (AM/AM and AM/PM) of Power Amplifiers (PAs). The implementation is realized in the Xilinx ISE tool with the Virtex-6 FPGA ML 605 Evaluation Kit using Verilog language. Experimental results have shown a high correlation with the inverse model computed with SystemVue in co-simulation with MATLAB for a GaN class F PA working with a LTE signal center at 2 GHz.

This paper presents a novel Field Programmable Gate Array (FPGA) architecture for hardware implementation of Multilayer Feedforward Neural Networks (MFNNs) suitable for Digital Pre-Distortion (DPD) technique. This architecture consists of a single neuron, several storage units and multiplexors that allow the reconfiguration via software of the FPGA to handle different numbers of inputs, layers, neurons and threshold functions of the MFNN. This novel FPGA architecture offers the advantage of using FPGAs with less logical resources than those found in a Virtex-6. The usefulness of this novel FPGA architecture is demonstrated by the modeling of the AM-AM and AM-PM characteristics of a GaN class F PA and a Doherty PA in a Virtex-6 FPGA from Xilinx configured with two different MFNNs, using a 2.1 GHz LTE signal of 5 MHz of bandwidth. The results obtained with the FPGA implementation are compared with data computed with MATLAB achieving Normalize Mean Square Error (NMSE) better than -50 dB.

One of the most common deficiencies in the public transport system is long waiting times. Currently, in the Guadalajara Metropolitan Area, Mexico, the frequencies of routes are fixed, making it impossible to satisfy a demand with a dynamic variation. An intelligent public transport system is required. The first step to solve this problem is knowing the number of users so that we can respond appropriately to each scenario. In this context, this work focuses on the design and implementation of an embedded system module for passenger counting that can be used to improves public transport service quality. This work presents three contributions. First, a design and experimental validation of the passenger counting system is presented to determine the number of users in an image and send this information to a server suitable for the public transportation system in Guadalajara, Mexico. Second, the generation of two new datasets is reported for training and testing the CSRNet algorithm with images of public transportation systems in Mexican cities. Finally, we make the hardware implementation of the passenger counting system in a Jetson Nano development board.

Lower-fidelity electromagnetic (EM) models are widely used to reduce the computational cost of microwave filter optimization, but their practical usefulness is often assessed through response similarity, own-fidelity objective values, or nominal simulation speedup. This paper examines whether designs optimized using coarse, mesh-derived EM models remain useful when verified at higher-fidelity. A cross-fidelity study is conducted using two microstrip filter benchmarks, a low-pass filter and a band-pass filter, under a fixed simulation setup, frequency sampling, objective function, and local-search procedure, so that mesh discretization is isolated as the source of model variation. Designs independently optimized at multiple mesh levels are reevaluated across the full model hierarchy, and their degradation during fine-model verification is used as an indicator of transferability. The results show that lower-fidelity success is not necessarily preserved under finer EM evaluation. In the band-pass filter case, extra coarse and coarse models satisfy the sampled transmission mask at their own-fidelity levels but yield substantially degraded responses in the fine model, whereas an intermediate model provides a more useful starting point for refinement. Space-mapping experiments further show that coarse-to-fine response matching does not always improve the target model objective value; its benefit depends on the selected model pair and on the observed transferability of the participating models. For the examples studied, the low-pass filter could be completed using an intermediate effective fine model, whereas the band-pass filter required an intermediate-to-fine progression followed by limited fine-model refinement. These findings highlight the need to evaluate coarse EM models by the survivability of their optimized designs, not only by speed or response similarity, before using them in multifidelity microwave filter optimization.