Search

In recent years, there has been a growing interest in using model-free deep reinforcement learning (DRL)-based controllers as an alternative approach to improve the dynamic behavior, efficiency, and other aspects of DC–DC power electronic converters, which are traditionally controlled based on small signal models. These conventional controllers often fail to self-adapt to various uncertainties and disturbances. This paper presents a design methodology using proximal policy optimization (PPO), a widely recognized and efficient DRL algorithm, to make near-optimal decisions for real buck converters operating in both continuous conduction mode (CCM) and discontinuous conduction mode (DCM) while handling resistive and inductive loads. Challenges associated with delays in real-time systems are identified. Key innovations include a chattering-reduction reward function, engineering of input features, and optimization of neural network architecture, which improve voltage regulation, ensure smoother operation, and optimize the computational cost of the neural network. The experimental and simulation results demonstrate the robustness and efficiency of the controller in real scenarios. The findings are believed to make significant contributions to the application of DRL controllers in real-time scenarios, providing guidelines and a starting point for designing controllers using the same method in this or other power electronic converter topologies.

The massive deployment of smart meters in most Western countries in recent decades has allowed the creation and development of a significant variety of applications, mainly related to efficient energy management. The information provided about energy consumption has also been dedicated to the areas of social work and health. In this context, smart meters are considered single-point non-intrusive sensors that might be used to monitor the behaviour and activity patterns of people living in a household. This work describes the design of a short-term behavioural alarm generator based on the processing of energy consumption data coming from a commercial smart meter. The device captured data from a household for a period of six months, thus providing the consumption disaggregated per appliance at an interval of one hour. These data were used to train different intelligent systems, capable of estimating the predicted consumption for the next one-hour interval. Four different approaches have been considered and compared when designing the prediction system: a recurrent neural network, a convolutional neural network, a random forest, and a decision tree. By statistically analysing these predictions and the actual final energy consumption measurements, anomalies can be detected in the undertaking of three different daily activities: sleeping, breakfast, and lunch. The recurrent neural network achieves an F1-score of 0.8 in the detection of these anomalies for the household under analysis, outperforming other approaches. The proposal might be applied to the generation of a short-term alarm, which can be involved in future deployments and developments in the field of ambient assisted living.

Training Deep Neural Networks to solve semantic segmentation is a challenging problem with small-size labeled datasets, leading to overfitting. This is especially problematic in medical images, and in particular, in laparoscopic surgery images. In this context, ground-truth segmentation labels are available only for a small set of images with few patients. Besides, inter-patient variability is very high in practice. Models trained for a specific setup and a set of patients usually performs poorly when deployed in a new environment. This work proposes a new training strategy that improves the generalization accuracy of current state-of-the-art semantic segmentation methods applied to laparoscopic images. Our approach is based on training a discriminator network, which learns to detect segmentation errors, producing a dense segmentation error map. Unlike in adversarial networks, we train the discriminator offline by synthetically altering ground-truth segmentation labels with simple morphological and geometric operations. We then use the discriminator to train a segmentation neural network, by minimizing the discriminator predicted error jointly with a standard segmentation loss. This strategy results in segmentation models that are significantly more accurate when tested in unseen images than those only relying on data augmentation. This technique is very suitable to boost the performance of any state-of-the-art segmentation network and can be combined with other data augmentation strategies. This paper evaluates and validates our proposal by training and testing common state-of-the-art segmentation models in publicly available semantic segmentation datasets, specialized in laparoscopic and endoscopic surgery. The results show that our methods are effective, obtaining a significant improvement in terms of segmentation accuracy, especially in challenging small-size datasets.

Shape-from-Template (SfT) solves the registration and 3D reconstruction of a deformable 3D object, represented by the template, from a single image. Recently, methods based on deep learning have been able to solve SfT for the wide-baseline case in real-time, clearly surpassing classical methods. However, the main limitation of current methods is the need for fine tuning of the neural models to a specific geometry and appearance represented by the template texture map. We propose the first texture-generic deep learning SfT method which adapts to new texture maps at run-time, without the need for texture specific fine tuning. We achieve this by dividing the problem into a segmentation step and a registration and reconstruction step, both solved with deep learning. We include the template texture map as one of the neural inputs in both steps, training our models to adapt to different ones. We show that our method obtains comparable or better results to previous deep learning models, which are texture specific. It works in challenging imaging conditions, including complex deformations, occlusions, motion blur and poor textures. Our implementation runs in real-time, with a low-cost GPU and CPU.

The research interest on location-based services has increased during the last years ever since 3D centimetre accuracy inside intelligent environments could be confronted with. This work proposes an indoor local positioning system based on LED lighting, transmitted from a set of beacons to a receiver. The receiver is based on a quadrant photodiode angular diversity aperture (QADA) plus an aperture placed over it. This configuration can be modelled as a perspective camera, where the image position of the transmitters can be used to recover the receiver’s 3D pose. This process is known as the perspective-n-point (PnP) problem, which is well known in computer vision and photogrammetry. This work investigates the use of different state-of-the-art PnP algorithms to localize the receiver in a large space of 2 × 2 m2 based on four co-planar transmitters and with a distance from transmitters to receiver up to 3.4 m. Encoding techniques are used to permit the simultaneous emission of all the transmitted signals and their processing in the receiver. In addition, correlation techniques (match filtering) are used to determine the image points projected from each emitter on the QADA. This work uses Monte Carlo simulations to characterize the absolute errors for a grid of test points under noisy measurements, as well as the robustness of the system when varying the 3D location of one transmitter. The IPPE algorithm obtained the best performance in this configuration. The proposal has also been experimentally evaluated in a real setup. The estimation of the receiver’s position at three particular points for roll angles of the receiver of γ={0°, 120°, 210° and 300°} using the IPPE algorithm achieves average absolute errors and standard deviations of 4.33 cm, 3.51 cm and 28.90 cm; and 1.84 cm, 1.17 cm and 19.80 cm in the coordinates x, y and z, respectively. These positioning results are in line with those obtained in previous work using triangulation techniques but with the addition that the complete pose of the receiver (x, y, z, α, β, γ) is obtained in this proposal.

A method for infrared and cameras sensor fusion, applied to indoor positioning in intelligent spaces, is proposed in this work. The fused position is obtained with a maximum likelihood estimator from infrared and camera independent observations. Specific models are proposed for variance propagation from infrared and camera observations (phase shifts and image respectively) to their respective position estimates and to the final fused estimation. Model simulations are compared with real measurements in a setup designed to validate the system. The difference between theoretical prediction and real measurements is between 0.4 cm (fusion) and 2.5 cm (camera), within a 95% confidence margin. The positioning precision is in the cm level (sub-cm level can be achieved at most tested positions) in a 4 × 3 m locating cell with 5 infrared detectors on the ceiling and one single camera, at distances from target up to 5 m and 7 m respectively. Due to the low cost system design and the results observed, the system is expected to be feasible and scalable to large real spaces.

This paper presents a novel approach for indoor acoustic source localization using microphone arrays, based on a Convolutional Neural Network (CNN). In the proposed solution, the CNN is designed to directly estimate the three-dimensional position of a single acoustic source using the raw audio signal as the input information and avoiding the use of hand-crafted audio features. Given the limited amount of available localization data, we propose, in this paper, a training strategy based on two steps. We first train our network using semi-synthetic data generated from close talk speech recordings. We simulate the time delays and distortion suffered in the signal that propagate from the source to the array of microphones. We then fine tune this network using a small amount of real data. Our experimental results, evaluated on a publicly available dataset recorded in a real room, show that this approach is able to produce networks that significantly improve existing localization methods based on SRP-PHAT strategies and also those presented in very recent proposals based on Convolutional Recurrent Neural Networks (CRNN). In addition, our experiments show that the performance of our CNN method does not show a relevant dependency on the speaker’s gender, nor on the size of the signal window being used.

This paper presents a sequential non-rigid reconstruction method that recovers the 3D shape and the camera pose of a deforming object from a video sequence and a previous shape model of the object. We take PTAM (Parallel Mapping and Tracking), a state-of-the-art sequential real-time SfM (Structure-from-Motion) engine, and we upgrade it to solve non-rigid reconstruction. Our method provides a good trade-off between processing time and reconstruction error without the need for specific processing hardware, such as GPUs. We improve the original PTAM matching by using descriptor-based features, as well as smoothness priors to better constrain the 3D error. This paper works with perspective projection and deals with outliers and missing data. We evaluate the tracking algorithm performance through different tests over several datasets of non-rigid deforming objects. Our method achieves state-of-the-art accuracy and can be used as a real-time method suitable for being embedded in portable devices.