Publications

Nowadays, one of the emergent challenges in mobile robotics consists of navigating safely and efficiently in dynamic environments populated by people. This paper focuses on the robot’s motion planning by proposing a learning-based method to adjust the robot’s trajectories to people’s movements by respecting the proxemics rules. With this purpose, we design a genetic algorithm to train the navigation stack of ROS during the goal-based navigation while the robot is disturbed by people. We also present a simulation environment based on Gazebo that extends the animated model for emulating a more natural human’s walking. Preliminary results show that our approach is able to plan people-aware robot’s trajectories respecting proxemics limits without worsening the performance in navigation.

Objective: The present investigation aims to compare the effect of a standardized exercise protocol in thermal aquatic immersion vs dryland (DL) on patients with chronic venous disease (CVD). Methods: Thirty-four patients with CVD (C3,Ep,As,Pr) were included in the study and randomly assigned to perform a standardized exercise protocol in a DL environment (DL group) or in a thermal water (TW group) from a natural hot spring at 33°C with a high mineral content. Leg volumetry, ankle range of motion (ROM), ultrasound-detected subcutaneous tissue and great saphenous vein (GSV) diameter were assessed. Quality of life was measured by VVSymQ and CIVIQ-20. Results: After five TW sessions the average volume decrease was –432.4 ± 122.4 mL (P < .0001) in the right leg and –358.8 ± 109.3 mL (P < .0001) in the left. No significant volume change was reported at the end of the five sessions in DL. In TW, the subcutaneous tissue thickness significantly decreased (all assessment points P < .0001 right and P < .0001 left). In contrast, no significant changes were found in the DL group. The TW group showed a significant great saphenous vein caliber reduction, both in the right and left legs (6.2 ± 5.9%, P < .002; 6.1 ± 2.2%, P < .0001), whereas in the DL group no significant differences were found. After five sessions, ankle ROM significantly increased in both groups, both in dorsiflexion and in plantarflexion (right leg: TWdorsiflexion P < .0001, TWplantarflexion P < .0001; DLdorsiflexion P < .003, DLplantarflexion P < .007) (left leg: TWdorsiflexion P < .0001, TWplantarflexion P < .0001; DLdorsiflexion P < .006, DLplantarflexion P < .001). Only the TW group showed a linear correlation between volume and ankle ROM variation (right leg: R2 = 0.80, R2 = 0.75, P < .0001; left leg: R2 = 0.82, R2 = 0.81, P < .0001). The VVSymQ and CIVQ20 scores significantly improved in TW (P < .0001 and P < .0001, respectively), whereas DL showed a significant improvement only in CIVQ20 score (P < .02). Conclusions: Thermal aquatic immersion enhances the clinical benefits of a standardized exercise protocol for patients with CVD. Compared with the data available in the literature on non-TW, the present investigation shows a potential role of higher density types of water in lower limb volume control. Intense and rigorous data collection is needed to move from empirical evidence to evidence-based science in TW, a potentially very useful treatment modality for CVD.

Falls due to balance impairment are a major cause of injury and disability in the elderly. The study of neurophysiological correlates during static and dynamic balance tasks is an emerging area of research that could lead to novel rehabilitation strategies and reduce fall risk. This review aims to highlight key concepts and identify gaps in the current knowledge of balance control in the elderly that could be addressed by relying on surface electromyographic (EMG) and electroencephalographic (EEG) recordings. The neurophysiological hypotheses underlying balance studies in the elderly as well as the methodologies, findings, and limitations of prior work are herein addressed. The literature shows: 1) a wide heterogeneity in the experimental procedures, protocols, and analyses; 2) a paucity of studies involving the investigation of cortical activity; 3) aging-related alterations of cortical activation during balance tasks characterized by lower cortico-muscular coherence and increased allocation of attentional control to postural tasks in the elderly; and 4) EMG patterns characterized by delayed onset after perturbations, increased levels of activity, and greater levels of muscle co-activation in the elderly compared to younger adults. EMG and EEG recordings are valuable tools to monitor muscular and cortical activity during the performance of balance tasks. However, standardized protocols and analysis techniques should be agreed upon and shared by the scientific community to provide reliable and reproducible results. This will allow researchers to gain a comprehensive knowledge on the neurophysiological changes affecting static and dynamic balance in the elderly and will inform the design of rehabilitative and preventive interventions.

Human re-identification is an important feature of domestic service robots, in particular for elderly monitoring and assistance, because it allows them to perform personalized tasks and human-robot interactions. However vision-based re- identification systems are subject to limitations due to human pose and poor lighting conditions. This paper presents a new re-identification method for service robots using thermal images. In robotic applications, as the number and size of thermal datasets is limited, it is hard to use approaches that require huge amount of training samples. We propose a re-identification system that can work using only a small amount of data. During training, we perform entropy-based sampling to obtain a thermal dictionary for each person. Then, a symbolic representation is produced by converting each video into sequences of dictionary elements. Finally, we train a classifier using this symbolic representation and geometric distribution within the new representation domain. The experiments are performed on a new thermal dataset for human re-identification, which includes various situations of human motion, poses and occlusion, and which is made publicly available for research purposes. The proposed approach has been tested on this dataset and its improvements over standard approaches have been demonstrated.

Human re-identification is an important feature of domestic service robots, in particular for elderly monitoring and assistance, because it allows them to perform personalized tasks and human-robot interactions. However vision-based re- identification systems are subject to limitations due to human pose and poor lighting conditions. This paper presents a new re-identification method for service robots using thermal images. In robotic applications, as the number and size of thermal datasets is limited, it is hard to use approaches that require huge amount of training samples. We propose a re-identification system that can work using only a small amount of data. During training, we perform entropy-based sampling to obtain a thermal dictionary for each person. Then, a symbolic representation is produced by converting each video into sequences of dictionary elements. Finally, we train a classifier using this symbolic representation and geometric distribution within the new representation domain. The experiments are performed on a new thermal dataset for human re-identification, which includes various situations of human motion, poses and occlusion, and which is made publicly available for research purposes. The proposed approach has been tested on this dataset and its improvements over standard approaches have been demonstrated.

Sound perception is a fundamental skill for many people with severe sight impairments. The research presented in this article is part of an ongoing project with the aim to create a mobile guidance aid to help people with vision impairments find objects within an unknown indoor environment. This system requires an effective non-visual interface and uses bone-conduction headphones to transmit audio instructions to the user. It has been implemented and tested with spatialised audio cues, which convey the direction of a predefined target in 3D space. We present an in-depth evaluation of the audio interface with several experiments that involve a large number of participants, both blindfolded and with actual visual impairments, and analyse the pros and cons of our design choices. In addition to producing results comparable to the state-of-The-Art, we found that Fitts’s Law (a predictive model for human movement) provides a suitable metric that can be used to improve and refine the quality of the audio interface in future mobile navigation aids.

Human re-identification is an important feature of domestic service robots, in particular for elderly monitoring and assistance, because it allows them to perform personalized tasks and human-robot interactions. However vision-based re- identification systems are subject to limitations due to human pose and poor lighting conditions. This paper presents a new re-identification method for service robots using thermal images. In robotic applications, as the number and size of thermal datasets is limited, it is hard to use approaches that require huge amount of training samples. We propose a re-identification system that can work using only a small amount of data. During training, we perform entropy-based sampling to obtain a thermal dictionary for each person. Then, a symbolic representation is produced by converting each video into sequences of dictionary elements. Finally, we train a classifier using this symbolic representation and geometric distribution within the new representation domain. The experiments are performed on a new thermal dataset for human re-identification, which includes various situations of human motion, poses and occlusion, and which is made publicly available for research purposes. The proposed approach has been tested on this dataset and its improvements over standard approaches have been demonstrated.

Modern service robots are provided with one or more sensors, often including RGB-D cameras, to perceive objects and humans in the environment. This paper proposes a new system for the recognition of human social activities from a continuous stream of RGB-D data. Many of the works until now have succeeded in recognising activities from clipped videos in datasets, but for robotic applications it is important to be able to move to more realistic scenarios in which such activities are not manually selected. For this reason, it is useful to detect the time intervals when humans are performing social activities, the recognition of which can contribute to trigger human-robot interactions or to detect situations of potential danger. The main contributions of this research work include a novel system for the recognition of social activities from continuous RGB-D data, combining temporal segmentation and classification, as well as a model for learning the proximity-based priors of the social activities. A new public dataset with RGB-D videos of social and individual activities is also provided and used for evaluating the proposed solutions. The results show the good performance of the system in recognising social activities from continuous RGB-D data.

Human re-identification is an important feature of domestic service robots, in particular for elderly monitoring and assistance, because it allows them to perform personalized tasks and human-robot interactions. However vision-based re- identification systems are subject to limitations due to human pose and poor lighting conditions. This paper presents a new re-identification method for service robots using thermal images. In robotic applications, as the number and size of thermal datasets is limited, it is hard to use approaches that require huge amount of training samples. We propose a re-identification system that can work using only a small amount of data. During training, we perform entropy-based sampling to obtain a thermal dictionary for each person. Then, a symbolic representation is produced by converting each video into sequences of dictionary elements. Finally, we train a classifier using this symbolic representation and geometric distribution within the new representation domain. The experiments are performed on a new thermal dataset for human re-identification, which includes various situations of human motion, poses and occlusion, and which is made publicly available for research purposes. The proposed approach has been tested on this dataset and its improvements over standard approaches have been demonstrated.

The application of an innovative solidification/stabilization (S/S) process was investigated for the remediation of Pb contaminated soil. The performance of Pb stabilization was evaluated by comparing the use of calcium aluminate cement (CAC) and an alkali activated metakaolin binder vs the Ordinary Portland Cement (OPC). The phase composition of the stabilized products was investigated by XRD and correlated to the internal microstructure obtained by SEM-EDX imaging. Leaching tests were performed to ascertain the effectiveness of the proposed binders in the S/S of the contaminated soil, and Pb release was evaluated for each binding system. The overall results proved that multiple mechanisms are involved in Pb retention and that key parameters regulating the stabilization performance are strongly dependent on the type of applied binder system. Pb was found to be associated to C-S-H in the case of OPC, whereas ettringite played a key role in the retention of this contaminant using the CAC binder. The use of a NaOH activated metakaolin resulted in almost total retention of Pb, despite a lack of solidification, highlighting the importance of pH in the regulation of the leaching behavior.