Emotion Recognition (ER) has gained significant attention due to its importance in advanced human-machine interaction and its widespread real-world applications. In recent years, research on ER systems has focused on multiple key aspects, including the development of high-quality emotional databases, the selection of robust feature representations, and the implementation of advanced classifiers leveraging AI-based techniques. Despite this progress in research, ER still faces significant challenges and gaps that must be addressed to develop accurate and reliable systems. To systematically assess these critical aspects, particularly those centered on AI-based techniques, we employed the PRISMA methodology. Thus, we include journal and conference papers that provide essential insights into key parameters required for dataset development, involving emotion modeling (categorical or dimensional), the type of speech data (natural, acted, or elicited), the most common modalities integrated with acoustic and linguistic data from speech and the technologies used. Similarly, following this methodology, we identified the key representative features that serve as critical emotional information sources in both modalities. For acoustic, this included those extracted from the time and frequency domains, while for linguistic, earlier embeddings and the most common transformer models were considered. In addition, Deep Learning (DL) and attention-based methods were analyzed for both. Given the importance of effectively combining these diverse features for improving ER, we then explore fusion techniques based on the level of abstraction. Specifically, we focus on traditional approaches, including feature-, decision-, DL-, and attention-based fusion methods. Next, we provide a comparative analysis to assess the performance of the approaches included in our study. Our findings indicate that for the most commonly used datasets in the literature: IEMOCAP and MELD, the integration of acoustic and linguistic features reached a weighted accuracy (WA) of 85.71% and 63.80%, respectively. Finally, we discuss the main challenges and propose future guidelines that could enhance the performance of ER systems using acoustic and linguistic features from speech. • Exploration of emotion modeling theories. • Analysis of key insights of Speech Emotion Recognition (SER) datasets. • Assessment of speech features across acoustic and linguistic modalities for ER. • Thorough evaluation of fusion techniques for SER. • Detailed benchmark analysis of SER approaches. • Detailed selection of research articles under systematic review approach. • Critical discussion of the current state and emerging research opportunities.

The scarcity of medical data, particularly in Survival Analysis (SA) for cancer-related diseases, challenges data-driven healthcare research. While Synthetic Tabular Data Generation (STDG) models have been proposed to address this issue, most rely on datasets with abundant samples, not reflecting real-world limitations. We suggest using an STDG approach that leverages transfer learning and meta-learning techniques to create an artificial inductive bias, guiding generative models trained on limited samples. Initial experiments were conducted on larger classification datasets, allowing us to asses the methodology across varying sample sizes and abundant versus scarce data scenarios. We primarily employed clinical utility validation for cancer-related SA data, as divergence-based similarity validation was not feasible. The methodology improved STDG under constrained data conditions, with divergence-based similarity validation proving to be a robust measure of data quality. Conversely, clinical utility validation yielded similar results regardless of sample size, indicating its limitations in statistically confirming effective STDG. In SA experiments, we observed that altering the task can reveal if relationships among variables are accurately generated, with most cases benefiting from the proposed methodology. Our study underscores the efficacy of the approach in tackling medical data scarcity by effectively generating high-quality synthetic data under constrained conditions. While divergence-based similarity validation is essential when sufficient data are available, clinical utility validation alone is insufficient and should be complemented by similarity validation. These findings underscore the potential and limitations of STDG methodologies in addressing medical data scarcity.

This paper presents the implementation of a localization system in a senior living residence located in a town in the province of Guadalajara, Spain. The system integrates precise location tracking using Ultra-Wideband (UWB) technology in the common areas of the residence, and symbolic positioning via millimeter-wave (mmWave) radar modules installed in private rooms. In UWB-monitored areas, users carry a wearable UWB tag, while in mmWave-radar monitored rooms, a device-free approach is adopted. The localization system is designed to be scalable, easy to deploy, and to support remote maintenance and management, facilitating long-term operation with minimal on-site intervention. This paper provides a detailed description of the devices used, the deployment process, and the overall system setup. Furthermore, by leveraging open-source tools, efficient data management has been achieved. Finally, raw data collected from the implemented system are presented graphically. These data will support future development of tools to detect behavioral anomalies early, which may indicate physical or cognitive decline among the monitored individuals. The collected data are also valuable for identifying daily behavioral patterns and routines, and for detecting deviations that could signal the onset of conditions associated with physical and/or cognitive impairment.

While synthetic tabular data generation using Deep Generative Models (DGMs) offers a compelling solution to data scarcity and privacy concerns, their effectiveness relies on the availability of substantial training data, often lacking in real-world scenarios. To overcome this limitation, we propose a novel methodology that explicitly integrates artificial inductive biases into the generative process to improve data quality in low-data regimes. Our framework leverages transfer learning and meta-learning techniques to construct and inject informative inductive biases into DGMs. We evaluate four approaches (pre-training, model averaging, Model-Agnostic Meta-Learning (MAML), and Domain Randomized Search (DRS)) and analyze their impact on the quality of the generated text. Experimental results show that incorporating inductive bias substantially improves performance, with transfer learning methods outperforming meta-learning, achieving up to 60% gains in Jensen-Shannon divergence. The methodology is model-agnostic and especially relevant in domains such as healthcare and finance, where high-quality synthetic data are essential, and data availability is often limited.

This work proposes the use of room-level or symbolic localisation for the analysis of the daily life of elderly people. The elderly people tend to be quite routine, so variations in behaviour patterns can be an indication of physical or cognitive impairment. Then, it is essential to develop solutions for the early detection of changes or anomalies that can be the onset of physical or cognitive impairment. In this work, the behaviour of the monitored person is modelled by the probability of staying in each room of their place of residence. The proposed system is based on the generation of a baseline behaviour, defined by groups of similar days during an initial training period. From this model, each new day is compared with the baseline behaviour and the evolution is analysed. Two databases are used for the evaluation of the proposed system: a real database and a synthetic one. The initial results are very promising, being possible to classify location-based daily activities as main and secondary routines, detecting when a day is considered unusual.

Gait analysis is the systematic study of the movement of human locomotion. It has applications in the field of health by allowing the diagnosis of diseases related to gait disturbances, such as the risk of falling or frailty in the elderly. The identification of these pathologies is carried out by analyzing different specific gait parameters, such as speed, time or length of each step, etc. Currently, there are numerous methods or technologies for gait analysis, mainly divided into non-portable systems (optical systems or pressure mapping corridors) and portable systems (inertial sensors, pressure insoles, electromyography, etc.), with different characteristics and limitations. It is important to be able to assess which technique is the most appropriate to use for each type of gait analysis required. In this work, we study and compare three different technologies for gait analysis, two of them commercial: an optical system (optitrack) and a pressure mapping system for human gait analysis (StrideWay); and a dual inertial system to be placed on the feet (IMUE-CSIC, proprietary design). We analyze the main features of each system, as well as their advantages and limitations, including the gait parameters that they can provide directly or indirectly from the recorded information. In addition, gait trials have been carried out using all three systems simultaneously. The results obtained reveal very small differences between them, being less than 0.02 seconds for gait cycle time and 0.03 meters for stride length. The advantages of using IMUs are their low cost and portability (allowing outdoor testing). Despite requiring complex post-processing algorithms, they provide a broad set of gait parameters. The optical system and the pressure mapping corridor for human gait analysis have high resolution, but their high cost and lack of portability make them difficult to use in real conditions outside the laboratory. Additionally, the pressure corridor provides more extensive results on the footprint, useful in physiotherapy or podiatry applications.

Home-based physical therapies are effective if the prescribed exercises are correctly executed and patients adhere to these routines. This is specially important for older adults who can easily forget the guidelines from therapists. Inertial Measurement Units (IMUs) are commonly used for tracking exercise execution giving information of patients’ motion data. In this work, we propose the use of Machine Learning techniques to recognize which exercise is being carried out and to assess if the recognized exercise is properly executed by using data from four IMUs placed on the person limbs. To the best of our knowledge, both tasks have never been addressed together as a unique complex task before. However, their combination is needed for the complete characterization of the performance of physical therapies. We evaluate the performance of six machine learning classifiers in three contexts: recognition and evaluation in a single classifier, recognition of correct exercises, excluding the wrongly performed exercises, and a two-stage approach that first recognizes the exercise and then evaluates it. We apply our proposal to a set of 8 exercises of the upper-and lower-limbs designed for maintaining elderly people health status. To do so, the motion of 30 volunteers were monitored with 4 IMUs. We obtain accuracies of 88 . 4 % and the 91 . 4 % in the two initial scenarios. In the third one, the recognition provides an accuracy of 96 . 2 %, whereas the exercise evaluation varies between 93 . 6 % and 100 . 0 %. This work proves the feasibility of IMUs for a complete monitoring of physical therapies in which we can get information of which exercise is being performed and its quality, as a basis for designing virtual coaches.

This work proposes a system that allows estimating and monitoring daily routine changes in a sensorized home through Machine Learning and Affinity Propagation clustering techniques. Older adults often have low-activity and rather routine lives, which means that these routines can be an indicator of their physical and cognitive state in order to lead an independent life and healthy ageing. Therefore, it is important to be able to generate precise routines, as well as to monitor them, to trigger alarms in case of significant variations. This proposal defines routines based on the time spent in each of the monitored rooms. The daily time in each room is estimated trough a Bluetooth Low Energy-based indoor localization system. The localization is obtained through the Bluetooth received signal strength, which is processed with different supervised algorithms and fused with the acceleration measured by the mobile receiver, obtaining an accuracy above 96 %. From these data, the sample has been synthetically expanded to generate four different routines, on which the proposed algorithm based on Principal Component Analysis and Affinity Propagation clustering has been tested, obtaining very promising results.

The main limitation of using inertial measurement units (IMUs) for inertial navigation systems (INS) is the accumulated error in the heading angle estimation. Different approaches have been proposed to reduce the position error caused by the heading drift, but accurate biomechanical models have not been explored in INS. A precise leg characterization method is needed to develop INS-oriented biomechanical models. The main goal of this work is to evaluate an IMU-based method to locate the leg joint centers and axes, and to estimate the leg segment lengths. These leg parameters are required for the implementation of lower-limb biomechanical models. Four different versions of the method are implemented and compared using a stereo- photogrammetric system as gold standard. The method validation is conducted in four volunteers by doing five exercises. The calibration exercises involve five different leg movements. The proposed method shows an average accuracy of 2 cm and 3 cm over 21cm and 54 cm in joint centers and axes determination, respectively. In the estimation of the leg segment lengths, the average accuracy of the proposed method is 2 cm over 39 cm. The proposed method brings promising results improving the estimations of the state-of-the-art methods.

This work proposes an Android-based platform to warn the medical staff of the onset of allergy reactions during a provocation test in a hospital. The portable system carries out the analysis of the heart rate variability for the early detection of allergic reactions in patients undergoing allergy provocation tests at hospitals. The proposal is composed of an ECG (electrocardiogram) acquisition system and an Android device (Smartphone, Tablet) that monitors and evaluates the results in real time, increasing the safety of allergic tests. At present, food and drug allergic tests are a major problem for patients because of their long duration and intrusion. However, the authors have designed an algorithm for detecting allergy reactions that have focused on reducing the time of the tests and the number of doses. This algorithm runs on an Android platform, and it is able to provide alarms for the medical staff if there is an allergy reaction. The proposed monitoring system is very suitable for the health monitoring during the provocation tests.

The current stage of railway transportation systems must deal with increased safety and reliability issues. A key point is improving occupancy track circuit performance and providing them with redundancy, higher noise immunity, and the capability to acquire additional information about the track section involved. This work proposes a novel track circuit based on the encoding of the electrical transmissions with Kasami codes. Track circuit emitters send signals coded with a known sequence that can be identified by the corresponding receivers using correlation techniques; these processes increase immunity to noise and changes in environmental conditions. An appropriate selection of orthogonal sequences for encoding, as well as different carrier frequencies for transmissions, allow simultaneous emissions and receptions without cross interference.

The need for monitoring underwater sensors, communications between submarines or sonar, makes underwater acoustic communications an important field of research. Electromagnetic waves are quickly attenuated in this medium and thus acoustic waves are the best option for communications. When two buoys are present, one acting as the emitter and the other as the receiver, it is crucial for the communication system to know the response of the channel. This paper presents a study of the underwater channel and proposes a model developed in Matlab where physical phenomena like attenuation, absorption and the swell effect are considered. This latter effect causes a Doppler spread in the signal, and a time-varying impulse response.

A method to locate reflectors, based on the spectral analysis of ultrasonic received echoes, is provided in this study. The main contribution of this work lies in the spectral analysis for bearing estimation. A single ultrasonic transducer, acting as emitter and receiver, is used. The transmitted signal is a 1023-bits Kasami code BPSK modulated by a carrier that ranges from 30 kHz to 55 kHz, implying a 6-component transmitted pulse. Since the ultrasonic transducer works as an angle-dependent filter, it is possible to estimate the bearing angle of the reflector based on the received spectrum. It is shown how the amplitude of each frequency contains information about the environment, particularly about the bearing angle of the reflector. Finally, a set of experimental results is included to validate the sensory system, showing how it can deal with realistic reflectors.

On current railway systems, it is becoming ever more necessary to install safety elements to avoid accidents. One of the causes that can provoke serious accidents is the existence of obstacles on the tracks, either fixed or mobile. In this paper, a multisensory system that can inform the monitoring system about the existence of obstacles is proposed. The system for obstacle detection consists of two emitting and receiving barriers, which are placed on opposing sides of the railway, respectively, and use infrared and ultrasonic sensors, thus establishing different optical and acoustic links between them. The interruption of one or several links should produce an alarm. However, even without the existence of objects, degradation of links could occur due to atmospheric attenuation, solar radiation, etc., also producing an activation of the alarm system. Since detection is based on the lack of radiation in the detectors, the use of complementary sensors for the same task is justified. Since the minimum size of an object for which an alarm is required to be generated is 50 × 50 × 50 cm, in some situations, several links are interrupted; however, alarms should not be generated. Typical cases are the flight of leaves or the movement of small animals in the scanned area. To avoid alarm activation in such situations, this paper proposes the combined use of diverse techniques of data fusion, based on fuzzy logic and the Dempster-Shafer theory of evidence, to validate the existence of objects, providing a highly reliable detection system.

Ultra-wideband (UWB) technology is currently considered the most attractive solution to develop a Local Positioning System. In fact, the IEEE 802.15.4a standard includes a link (MAC) and physical (PHY) layer with indoor positioning capabilities. This paper presents a study of the typical modulations schemes used with UWB and its adaptation to ultrasonic frequencies. Moreover, several tests of the modulation schemes in the ultrasonic frequency range are shown, in order to apply these techniques in a local positioning system. The main goal is to implement the algorithms in FPGAs using UWB radio technology. Therefore, the use of ultrasonic frequencies can be considered as a previous step of the future work.

The current development of railway transport gives a significant importance to the design of suitable and reliable safety systems. Among them, the common track circuit can be improved to enhance its performance, by providing redundancy, immunity to noise and more information from the track circuit. This work proposes a novel track circuit based on the encoding of the electrical transmissions with Kasami codes. Every node emits its corresponding code that can be detected at next nodes by correlation techniques. The suitable selection of orthogonal sequences, as well as different carrier frequencies for emissions, allow to have simultaneous emissions and receptions without cross interferences.

Loosely Synchronous (LS) codes improves the performance of Local Positioning Systems (LPS) in terms of noise immunity, capability of simultaneous measurements and precision in the location, provided that the multipath spread and the relative time offset between the codes are within the zero correlation zone (ZCZ) that they achieve in their correlation functions. Furthermore, these codes can be very efficiently generated and correlated if compared to a straightforward implementation. This work presents the hardware implementation of the correlation tasks associated to an ultrasonic LPS based on LS sequences generated from Complementary Sets of Sequences (CSS). This implementation is able to carry out all the correlation processing tasks in a time shorter than the sampling period.

In this work the specific software tools designed to manage a high performance data acquisition board to be used in active multi-sensory systems is presented. Specifically, the developments which has been carried out allow to monitor the coded emission/reception of two sensory barriers for obstacle detection on railways, based on infrared (IR) and ultrasonic (US) sensors. Additionally, it has been developed the tools to detect the coded emissions. As a result of that, it can be detected the existence of obstacles between both barriers, based on the lack of signal at the receiving barrier. The developed software can be considered the previous step to implement the hardware prototype. Initially it has been validated the feasibility of the algorithm, optimizing the processing time to obtain the detection results in quasi-real time.

In local positioning systems (LPS) active beacons are placed in the environment, so the accurate coordinates of these beacons are necessary for positioning algorithms. Such coordinates or positions can be obtained by means of hand-made measurements (a slow and less flexible method), or by using calibration algorithms (i.e., automatic determination of the beacon coordinates from pre-determined measurements). In this work a new method is presented to estimate the optimal position of the test-points, necessary to calibrate an ultrasonic LPS. The method has been developed for both, spherical and hyperbolic trilateration. The results from simulated and real data shows the improvement obtained when using in the calibration these optimal test-points instead of randomly selected ones.

The problem of database integration has been widely tackled through different approaches. While data transformation based systems, such as Data Warehouses, reached the acceptation of the industry during the 80's, in the last decade query translation based approaches have gained popularity given their adequacy to dynamic domains. While the former are based on gathering actual data in central repositories, the latter allow data to remain in the original databases. There still exist several issues to be tackled in query translation, mainly if no ad-hoc schema is described, such as problems with scalability and query processing. In this paper we describe a complete database integration and semantic mediation approach, borrowing techniques from both data transformation and local as view data translation methods, and addressing the cross referencing problem. This work has been carried out in the framework of ACGT (Advancing Clinico-Genomic Trials on Cancer) project, supported by the European Commission.

The determination of the relative position among mobile objects or members of a robot team is a useful information when the systems are not restricted to a particular environment and they cannot depend on an external infrastructure to obtain their location information. The solution for this kind of problems cannot be divided and treated separately for each object. Therefore it is necessary to implement mechanisms that allow to simultaneously determine the spatial relations among objects not knowing their absolute location in the environment. After that, a positioning algorithm can be computed whose complexity depends on the number of observations made and the precision required in the system. In this work, the implementation of a relative positioning system is described where only acoustic emissions are made in order to obtain information about the other objects. These data are obtained by using simultaneous round-trip-time-of-flight measurements. Finally, using this information, the object positions are computed with a multidimensional scaling technique, and a closer solution is achieved with least-square algorithms.

Kasami sequences have been successfully used in communications, navigation and related systems due to their low cross-correlation values, compared to those from other binary sequences. In this work, different alternatives for the hardware implementation of a correlator of Kasami sequences are presented: for short sequences a combinational design is proposed, whereas three sequential designs are suggested for longer Kasami sequences. These three sequential designs differ about the management of the memory: one stores the necessary data in slices of the FPGA; another uses external memory; and finally, the last one uses the internal RAM blocks in the FPGA.

Signal coding and pulse compression techniques have been widely used in the design of high performance airborne sonar as a mean to increase their robustness to noise and enhance their spatial resolution. These systems are still today in continuous development, with the appearance of new encoding schemes and processing algorithms which not only improve the previous features, but also provide the system with additional properties. All these improvements are achieved at the expense of increasing the computational complexity of the signal processing tasks, what put into risk the real-time operation capability of the system. This work presents the hardware implementation in a single FPGA of all the stages forming the detection module of an advanced sonar system which is based on the emission of complementary sets of sequences. This implementation is able to carry out all the data processing in the time interval between two successive firings, also obtaining a portable module that makes possible its incorporation into mobile robots