1,721,088 research outputs found
Path Planning and Following for an Autonomous Model Car Using an “Eye in the Sky”
No full textAutonomous driving is a trend topic that is enabled by communication between devices. Under this scope, ROS is a useful tool for running multiple processes in a graph architecture, where each node may receive and post messages that are consumed by other nodes for their own needs. In this tutorial chapter we discuss a solution for autonomous path planning using a Randomized Random Tree (RRT) and a simple control scheme based on PIDs to follow that path. The control uses internal sensors and an external camera that works as an “eye in the sky”. This is implemented with the help of ROS version 1.12.13 using the Kinetic distribution. Results are validated using the Gazebo multi-robot simulator, version 7.0.0. The robot model used corresponds to the AutoNOMOS mini developed by PHD Raúl Rojas, while the “eye in the sky” is an artificial simple RGB camera created in Gazebo for research purposes. The Rviz package is used to monitor the simulation. The repository for this project can be found at https://github.com/Sutadasuto/AutoNOMOS_Stardust. (The original model for the AutoNOMOS mini was retrieved from https://github.com/EagleKnights/EK_AutoNOMOS_Sim ).</p
Parametric Optimization for Nonlinear Quadcopter Control Using Stochastic Test Signals
No full textA key activity in the deployment of quadcopters is controller tuning. This research chapter addresses the problem of how to optimize the parameter set of a controller for a quadcopter. Existing research in iterative controller optimization has centered on the use of linear models of the process. However, in this research chapter, we propose a procedure based on conjugate gradient optimization for controller tuning when the dynamic model is nonlinear and the test signals are stochastic. To validate the findings, a bipartite ROS application was implemented. The first part corresponds to the orientation controller of the drone which runs on the onboard computer. The second part carries out the position controller and runs on a ground station computer. ROS Indigo Igloo is used for the code of this chapter.</p
SkiROS—A skill-based robot control platform on top of ROS
No full textThe development of cognitive robots in ROS still lacks the support of some key components: a knowledge integration framework and a framework for autonomous mission execution. In this research chapter, we will discuss our skill-based platform SkiROS, that was developed on top of ROS in order to organize robot knowledge and its behavior. We will show how SkiROS offers the possibility to integrate different functionalities in form of skill ‘apps’ and how SkiROS offers services for integrating these skill-apps into a consistent workspace. Furthermore, we will show how these skill-apps can be automatically executed based on autonomous, goal-directed task planning. SkiROS helps the developers to program and port their high-level code over a heterogeneous range of robots, meanwhile the minimal Graphical User Interface (GUI) allows non-expert users to start and supervise the execution. As an application example, we present how SkiROS was used to vertically integrate a robot into the manufacturing system of PSA Peugeot-Citroën. We will discuss the characteristics of the SkiROS architecture which makes it not limited to the automotive industry but flexible enough to be used in other application areas as well. SkiROS has been developed on Ubuntu 14.04 LTS and ROS indigo and it can be downloaded at https://github.com/frovida/skiros. A demonstration video is also available at https://youtu.be/mo7UbwXW5W0.</p
Video Stabilization of the NAO Robot Using IMU Data
No full textThe implementation of a video stabilization system of the NAO robot is presented through the data of the IMU, this stabilized image is used to detect and track QR codes. Once the QR code is located, the NAO robot tracks and monitors it. The system was developed under the ROS platform, with modules implemented in C++ and Python languages. The system provides data for subsequent processes, which need to use video data for object recognition, task tracking, among others. Can get sequences of stable images and with the least amount of vibrations or sudden movements. One of the main benefits of this work is the visual tracking of objects through stable images during walking of the NAO robot, which introduces an erratic motion of the head camera, the effect that is mitigated with the digital visual gyrostabilized method presented in this work.</p
Unmanned Aerial Systems : Autonomy, Cognition and Control
No full textIncreasing trend towards higher level of autonomy in unmanned aerial systems (UASs) requires less control by the human operator and increasing capability to perform complex tasks by reacting to the environmental influences. Nevertheless, current UASs, are designed to function in static, and predictable environments. Therefore, it is envisaged that the existing uncertainties and dynamic changes, caused when an unmanned aerial vehicle (UAV) is operating in an unknown environment, would degrade its performance signicantly. The uncertainties can be also incurred through interaction with other complex and intelligent systems, such as humans. We present a compact literature survey of UASs control and navigation as a basic knowledge to develop UASs from the perspective of control engineer. Besides, we present several control strategies to maintain a UAS, as well as multi-UASs under a network setting under various scenarios. Several simulations are given to illustrate the performance of the controllers in MATLAB. Advances in computing power and algorithms currently enable development of systems with high degree of autonomy. Nonetheless, there is a large gap between practical operation in a real-world and laboratory implementation, as safe deployment of UASs, requires validation of their behaviour under almost all envisaged scenarios. A reliable and autonomous operation of such a system requires design and development of a cognitive control system that acquires knowledge and understanding of the surrounding environment via perception, reasoning and learning. Cognitive control systems in UASs will enhance their safety and performance. Cognitive control can also be used in cooperative execution of complex tasks where multiple agents such as humans, machines or both interact. Such UASs will have a great potential to be used in extreme environments such as search and rescue in case of disaster, nuclear decommissioning operation, deep-sea exploration, mining, etc
Containerization: For Over-the-Air Programming of Field Deployed Internet-of-Energy Based on Cost Effective LPWAN
No full textContainerization is widely investigated as a secure and lightweight virtualization solution. They have outclassed the traditional virtual machines (VMs) architecture in the cloud because of having built-in capabilities to provide platform-as-a-service (PaaS) to the edge devices. Containers consume less hardware resource, scalable and provide faster patch times, thus they have been suggested as a solution for more practical and real-time virtualization resource in energy utilities. Containers can be used as a replacement for VMs because the main feature of the modern cloud architecture, i.e. workloads are able to utilize shared hardware resources that are provided across all data-center nodes for any purpose. This article reviews the implementation architecture of the containers for energy sector applications, i.e. next-generation substation automation systems. Moreover, we analysed the power consumed by the edge devices, based on the low power wide area network (LPWAN), for over-the-air programming through containers. In addition, we studied the success factors of containers, by considering specific used-cases and explain the benefits of using containers over traditional VMs
HyperFlex: a model driven toolchain for designing and configuring software control systems for autonomous robots
No full textA huge corpus of open source robotic software libraries is available on ROS repositories that can be reused to develop a large variety of robot control systems. The difficult challenge consists in selecting and integrating a coherent set of components that provide the required functionality taking into account their mutual dependencies and architectural mismatches. The HyperFlex approach presented in this chapter enables the explicit representation of robot system architectures, functional variability, and application requirements as softwaremodels that can be manipulated by a system configuration engine
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Enhancing Speaker Diarization in Forensic Audio:A Comparative Analysis of Machine Learning Algorithms for Gender Classification
No full textSpeaker diarization is vital in contexts like police interrogations, where it enhances the security and personalization of data access and improves confidentiality in multi-speaker environments. The transcription of low-quality forensic audio recordings is challenging, as they are often marred by unclear speech and impede the accuracy of conventional Automatic Speech Recognition (ASR) systems. This paper evaluates the efficacy of traditional machine learning algorithms—Support Vector Machine (SVM), Decision Tree Classifier, Random Forest Classifier, and XGBoost in gender classification from voice samples for speaker diarization systems. These systems are critical in contexts like police interrogations, where they enhance data security and improve confidentiality in multi-speaker environments. We test these algorithms against real-world data, simulating practical conditions to ensure robustness. Our findings reveal that ensemble methods, particularly Random Forest and XGBoost, demonstrate high accuracy and strong generalizability when dealing with unfiltered, real-world audio data. XGBoost shows significant resistance to overfitting, making it highly suitable for secure voice-driven applications. This study aids in algorithm selection for speaker diarization tasks. It addresses gaps in forensic audio transcription accuracy, thereby enhancing the reliability of transcriptions and reducing risks of erroneous interpretations in legal contexts
- …
