Polytechnic Institute of Porto
Repositório Científico do Instituto Politécnico do PortoNot a member yet
26471 research outputs found
Sort by
On-Board Deep Q-Network for UAV-Assisted Online Power Transfer and Data Collection
No full textUnmanned Aerial Vehicles (UAVs) with Microwave
Power Transfer (MPT) capability provide a practical means to deploy a large number of wireless powered sensing devices into areas
with no access to persistent power supplies. The UAV can charge the
sensing devices remotely and harvest their data. A key challenge is
online MPT and data collection in the presence of on-board control
of a UAV (e.g., patrolling velocity) for preventing battery drainage
and data queue overflow of the devices, while up-to-date knowledge
on battery level and data queue of the devices is not available at
the UAV. In this paper, an on-board deep Q-network is developed
to minimize the overall data packet loss of the sensing devices,
by optimally deciding the device to be charged and interrogated
for data collection, and the instantaneous patrolling velocity of
the UAV. Specifically, we formulate a Markov Decision Process
(MDP) with the states of battery level and data queue length of
devices, channel conditions, and waypoints given the trajectory of
the UAV; and solve it optimally with Q-learning. Furthermore, we
propose the on-board deep Q-network that enlarges the state space
of the MDP, and a deep reinforcement learning based scheduling
algorithm that asymptotically derives the optimal solution online,
even when the UAV has only outdated knowledge on the MDP
states. Numerical results demonstrate that our deep reinforcement
learning algorithm reduces the packet loss by at least 69.2%, as
compared to existing non-learning greedy algorithms.info:eu-repo/semantics/publishedVersio
Car Detection using Unmanned Aerial Vehicles: Comparison between Faster R-CNN and YOLOv3
No full textUnmanned Aerial Vehicles are increasingly being used in surveillance and traffic monitoring thanks to their high mobility and ability to cover areas at different altitudes and locations. One of the major challenges is to use aerial images to accurately detect cars and count-them in real-time for traffic monitoring purposes. Several deep learning techniques were recently proposed based on convolution neural network (CNN) for real-time classification and recognition in computer vision. However, their performance depends on the scenarios where they are used. In this paper, we investigate the performance of two state-of-the art CNN algorithms, namely Faster R-CNN and YOLOv3, in the context of car detection from aerial images. We trained and tested these two models on a large car dataset taken from UAVs. We demonstrated in this paper that YOLOv3 outperforms Faster R-CNN in sensitivity and processing time, although they are comparable in the precision metric.info:eu-repo/semantics/publishedVersio
Activity Monitoring of Islamic Prayer (Salat) Postures using Deep Learning
No full textIn the Muslim community, the prayer (i.e. Salat) is the second pillar of Islam, and it is the most essential and fundamental worshiping activity that believers have to perform five times a day. From a gestures' perspective, there are predefined human postures that must be performed in a precise manner. However, for several people, these postures are not correctly performed, due to being new to Salat or even having learned prayers in an incorrect manner. Furthermore, the time spent in each posture has to be balanced. To address these issues, we propose to develop an artificial intelligence assistive framework that guides worshippers to evaluate the correctness of the postures of their prayers. This paper represents the first step to achieve this objective and addresses the problem of the recognition of the basic gestures of Islamic prayer using Convolutional Neural Networks (CNN). The contribution of this paper lies in building a dataset for the basic Salat positions, and train a YOLOv3 neural network for the recognition of the gestures. Experimental results demonstrate that the mean average precision attains 85% for a training dataset of 764 images of the different postures. To the best of our knowledge, this is the first work that addresses human activity recognition of Salat using deep learning.info:eu-repo/semantics/publishedVersio
Building a Text Messaging-Based System to Support Low-Cost Automation in Household Agriculture
No full textHome garden crops, small-scale agricultural systems for local food production, are becoming an integral part of the food supply chain in a number of developing countries and peri-urban areas of some developed regions. In this work, we propose a low-cost, monitoring and irrigation system which can be applied in household and local community gardens. The basic architecture consists of a sensing/actuation station based on commercial off-the-shelf hardware and a mobile application for the interaction with remote users. A key aspect of the system is the use of legacy text-messaging service as a mechanism to support alert and control operations for monitoring and irrigation. This feature enables widely available and highly-reliable connections between the cropland station and remote users without the need for new network infrastructure. We implemented a functional prototype of the system to check its effectivity in a small open-field area for tomatoes cultivation. The results show that water usage can be substantially improved if using both the actual information collected from the system and public tools for decision support in agriculture. We conclude that the proposed solution has a good prospect as an input for the design of more automated decision-strategies to be used in plant cultivation of a similar kind and/or of a larger scale.info:eu-repo/semantics/publishedVersio
The Complete Reference (Volume 5)
No full textThis book is the fifth volume in the successful book series Robot Operating System: The Complete Reference.
The objective of the book is to provide the reader with comprehensive coverage on the Robot Operating System (ROS), which is currently considered to be the primary development framework for robotics applications, and the latest trends and contributing systems.
The content is divided into six parts. Pat I presents for the first time the emerging ROS 2.0 framework, while Part II focuses on multi-robot systems, namely on SLAM and Swarm coordination. Part III provides two chapters on autonomous systems, namely self-driving cars and unmanned aerial systems. In turn, Part IV addresses the contributions of simulation frameworks for ROS. In Part V, two chapters explore robotic manipulators and legged robots. Finally, Part VI presents emerging topics in monocular SLAM and a chapter on fault tolerance systems for ROS. Given its scope, the book will offer a valuable companion for ROS users and developers, helping them deepen their knowledge of ROS capabilities and features.info:eu-repo/semantics/publishedVersio
COPADRIVe - A Realistic Simulation Framework for Cooperative Autonomous Driving Applications
No full textSafety-critical cooperative vehicle applications such as platooning, require extensive testing, however, the complexity and cost involved in this process, increasingly demands for realistic simulation tools to ease the validation of such technologies, helping to bridge the gap between development and real-word deployment. In this paper we propose a realistic co-simulation framework for cooperative vehicles, that integrates Gazebo, an advanced robotics simulator, with the OMNeT++ network simulator, over the Robot Operating System (ROS) framework, supporting the simulation of advanced cooperative applications such as platooning, in realistic scenarios.info:eu-repo/semantics/publishedVersio
Sail Car—An EPS©ISEP 2019 Project
No full textThis paper provides an overview of the development of a Sail Car within the European Project Semester (EPS), the international multidisciplinary engineering capstone programme offered by the Instituto Superior de Engenharia do Porto (ISEP). The main goal of EPS@ISEP is to offer a project-based educational experience to develop teamwork, communication, interpersonal and problem-solving skills in an international and multidisciplinary set up. The Sail Car team consisted of six Erasmus students, who participated in EPS@ISEP during the spring of 2019. The objective of the project was to design and develop a wind-powered, easy to drive land sailing vehicle. First, the team researched existing commercial solutions and considered the marketing, ethics and sustainability dimensions of the project. Next, based on these studies, specified the full set of requirements, designed the Sailo solution and procured the components and materials required to build a real size proof-of-concept prototype. Finally, the team assembled and tested successfully the prototype. At the end of the semester, the team considered EPS@ISEP a mind-opening opportunity.info:eu-repo/semantics/publishedVersio
Phase selective growth of Cu12Sb4S13 and Cu3SbS4 thin films by chalcogenization of simultaneous sputtered metal precursors
No full textIn this work, we present a procedure to grow Cu12Sb4S13 and Cu3SbS4 thin films consisting of the deposition of simultaneously sputtered metal precursors followed by a annealing treatment in a sulphur atmosphere. The selection of the ternary phase is performed by adjusting the sulphur evaporation temperature in the chalcogenization process. It is shown that for a sulphur evaporation temperature of 140 ∘C the predominant phase is Cu12Sb4S13 while for 180 ∘C the predominant phase is Cu3SbS4. In order to ensure precursor composition homogeneity, the Cu-Sb metallic precursors are deposited simultaneously by RF magnetron sputtering using adjustable segmented targets. The morphological characterization of the films was made by scanning electron microscopy and the composition was analysed by energy dispersive spectroscopy. The structural analysis and phase identification were performed by X-ray diffraction and Raman scattering. The optical properties were studied on films deposited directly on bare glass and the optical bandgap energies of 1.47 eV and 0.89 eV for Cu12Sb4S13 and Cu3SbS4, respectively, were determined.info:eu-repo/semantics/publishedVersio
Classification and Recommendation With Data Streams
No full textNowadays, with the exponential growth of data stream sources (e.g., Internet of Things [IoT], social networks, crowdsourcing platforms, and personal mobile devices), data stream processing has become indispensable for online classification, recommendation, and evaluation. Its main goal is to maintain dynamic models updated, holding the captured patterns, to make accurate predictions. The foundations of data streams algorithms are incremental processing, in order to reduce the computational resources required to process large quantities of data, and relevance model updating. This article addresses data stream knowledge processing, covering classification, recommendation, and evaluation; describing existing algorithms/techniques; and identifying open challenges.info:eu-repo/semantics/publishedVersio
On Feasibility of Multichannel Reconfigurable Wireless Sensor Networks Under Real-Time and Energy Constraints
No full textThis paper deals with the medium between two reconfigurable sensor nodes characterized by radio interfaces that support multiple channels for exchanging real-time messages under energy constraints. These constraints are violated if the consumed energy in transmission is higher than the remaining quantity of energy. A reconfiguration, i.e., any addition or removal of tasks in devices and consequently of messages on the medium, can cause the violation of real-time or energy constraints at run time. To achieve a feasible scheduling in time (i.e., message deadlines will be respected) and energy (i.e., there is available energy) on the medium, we propose new dynamic solutions: Balance, Dilute, and a Combination of them to manage any addition or removal of messages. The proposed approach utilizes the energy harvesting techniques and the PowerControl algorithm to reduce the nonharvested consumed energy. The proposed strategies achieve significant improvement over existing methods and provide the highest percentage of adding messages, with a lower average in response time and energy consumption. They reach a percentage of success in adding the highest priority messages while meeting deadlines up to 85%.info:eu-repo/semantics/publishedVersio