作者:
Issat, HazemTar, Jozsef K.Obuda Univ
Doctoral Sch Safety & Secur Sci Budapest Hungary Obuda Univ
Antal Bejczy Ctr Intelligent Robot ABC iRob Univ Res & Innovat Ctr Budapest Hungary
The limited output of various drives means a challenge in controller design whenever the acceleration need of the "nominal trajectory to be tracked" temporarily exceeds the abilities of the saturated control...
详细信息
ISBN:
(纸本)9781728173764
The limited output of various drives means a challenge in controller design whenever the acceleration need of the "nominal trajectory to be tracked" temporarily exceeds the abilities of the saturated control system. The prevailing control design methods can tackle this problem either in a single theoretical step or in two consecutive steps. In this latter case in the first step the design happens without taking into account the actuator constraints, then apply a saturation compensator if the phenomenon of windup is observed. In the Fixed Point Iteration-based Adaptive control (FPIAC) that has been developed as an alternative of the Lyapunov function-based approach the actuator saturation causes problems in its both elementary levels: in the kinematic/kinetic level where the desired acceleration is calculated, and in the iterative process that compensates the effects of modeling errors of the dynamic system under control and that of the external disturbances. The here presented approach tackles this problem in both levels by relatively simple considerations. To illustrate the method's efficiency simulation investigations were done in the FPIAC control of a modification of the van der Pol oscillator to which an additional strongly nonlinear term was added.
This study is motivated by evolutionary robot systems where robot bodies and brains evolve simultaneously. In such systems robot 'birth' must be followed by 'infant learning' by a learning method that ...
详细信息
Unmanned Aerial Vehicle (UAV) operators must maintain high levels of situation awareness on their area of operation. To achieve this, they use the a Command and control (C2) map, which is shared among forces, and regu...
详细信息
ISBN:
(纸本)9781728160016
Unmanned Aerial Vehicle (UAV) operators must maintain high levels of situation awareness on their area of operation. To achieve this, they use the a Command and control (C2) map, which is shared among forces, and regularly overloaded with data that is irrelevant to their operational mission. Operators require distilled information at the right timing. Yet, the existing filtering mechanisms for C2 maps are layer-based and insufficient. We propose a new approach to automatically and dynamically filter information items on the map based on environmental and mission context. To achieve this, we introduce a three-tiers artificial intelligence (AI) based algorithm (GiCoMAF), where we delineate the use of machine learning (ML) models to support UAV missions. For the GiCoMAF development, tagged data was collected in simulated experimental runs with professional UAS operators. Different types of ML models were evaluated and fitted into the algorithm. The models achieved a relatively high accuracy at modeling human preference and area of interest. The approach presented in this study can be further implemented to support time-critical spatial-temporal operational problems.
This work investigates the fault-tolerant control scheme for Takagi-Sugeno fuzzy semi-Markov jump systems with actuator faults. The hidden semi-Markov model is employed to describe the asynchronous phenomenon between ...
详细信息
ISBN:
(纸本)9781665438575
This work investigates the fault-tolerant control scheme for Takagi-Sugeno fuzzy semi-Markov jump systems with actuator faults. The hidden semi-Markov model is employed to describe the asynchronous phenomenon between the controller and plant. Based on Lyapunov theory, the fault tolerant controller depending on observable mode is presented to make sure that the semi-Markov jump systems is stochastically stable and satisfies the $H_{\infty}$ performance. A simulation example via a single-link robot is employed to prove the correctness of the studied approach.
Mobile users typically provide their credentials once for authentication purposes. However, as long as the user remains active in the system, there are no mechanisms to verify whether the user who provides the credent...
详细信息
ISBN:
(纸本)9781728190495
Mobile users typically provide their credentials once for authentication purposes. However, as long as the user remains active in the system, there are no mechanisms to verify whether the user who provides the credential is still in control of the device. In mobile device systems, the user's identity is checked only in the login stage. Continuous authentication based on the user's touch behavior patterns could be a remarkable solution to verify without disruption that a legitimate user is still in control of a mobile device. In this research, an Artificial Immune System (AIS) is proposed to continuously authenticate the users in the background based on the user's touch gestures. Experiments with Clonal Selection (CS) and Negative Selection (NS) are conducted on three datasets. Our results show that the best AIS performance is able to authenticate 99.89% of the users correctly.
In this paper, we propose an action detection model. a simple yet effective combination of channel attention mechanism with 3D convolution neural network, by which to enhance the performance of feature extraction in t...
详细信息
ISBN:
(纸本)9781728125473
In this paper, we propose an action detection model. a simple yet effective combination of channel attention mechanism with 3D convolution neural network, by which to enhance the performance of feature extraction in the video. Channel attention module uses the channel information in the feature extraction process of our 3D convolutional network to efficaciously pick out the features that are essential for the task and suppress useless ones. The proposed model can effectively promote the spatio-temporal feature representation power in the video. We embed our feature extraction model into the R-C3D framework to test the performance of our method by conducting comparative experiments on the THUMOS'14 dataset. Experimental results indicate that the proposed method can authentically enhance the accuracy of action detection.
Nowadays in road transport lower fuel consumption, higher traffic safety and reduced environmental impact are in the focus of the developments. To reach these future objectives it is necessary to increase the level of...
详细信息
ISBN:
(纸本)9781728156255
Nowadays in road transport lower fuel consumption, higher traffic safety and reduced environmental impact are in the focus of the developments. To reach these future objectives it is necessary to increase the level of automation of road vehicles. Driving a road vehicle by a software is a complex controlling task. This paper presents a basic longitudinal motion controller, which uses a Linear-Quadratic (LQ) regulator. This controller is widespread in the controlling of MIMO systems, because it provides an interactive control, optimized for certain elements of the state vector. The aim of this work is to create an operative longitudinal vehicle control based on an LQ regulator and use it in a demonstration vehicle.
The proceedings contain 419 papers. The topics discussed include: AiCE: automating horizon scanning for the detection of emerging technologies;DeepHealth: deep representation learning with autoencoders for healthcare ...
ISBN:
(纸本)9781728125473
The proceedings contain 419 papers. The topics discussed include: AiCE: automating horizon scanning for the detection of emerging technologies;DeepHealth: deep representation learning with autoencoders for healthcare prediction;simp3: social interaction-based multi-pedestrian path prediction by self-driving cars;exploring reward-based hyper-heuristics for the job-shop scheduling problem;data-driven neuro ARCH (DDNA) volatility model for option pricing on cloud resources;combined selection and parameter control of meta-heuristics;scale-invariance ideas explain the empirical soil-water characteristic curve;entropy-based recognition of anomalous answers for efficient grading of short answers with an evolutionary clustering algorithm;short-term trajectory planning in TORCS using deep reinforcement learning;and analyzing of LAM-CIoT: lightweight authentication mechanism in cloud-based IoT environment.
The varying system parameters, end effector payload and environmental uncertainties are quite natural in real-world robotics applications. Therefore in order to adapt to the changing control environment and improving ...
详细信息
This paper introduces SfmCAD, a novel unsupervised network that reconstructs 3D shapes by learning the Sketchbased Feature Modeling operations commonly used in modern CAD workflows. Given a 3D shape represented as vox...
详细信息
ISBN:
(数字)9798350353006
ISBN:
(纸本)9798350353013
This paper introduces SfmCAD, a novel unsupervised network that reconstructs 3D shapes by learning the Sketchbased Feature Modeling operations commonly used in modern CAD workflows. Given a 3D shape represented as voxels, SfmCAD learns a neural-typed sketch+path parameterized representation, including 2D sketches of feature primitives and their 3D sweeping paths without supervision, for inferring feature-based CAD programs. SfmCAD employs 2D sketches for local detail representation and 3D paths to capture the overall structure, achieving a clear separation between shape details and structure. This conversion into parametric forms enables users to seamlessly adjust the shape's geometric and structural features, thus enhancing interpretability and user control. We demonstrate the effectiveness of our method by applying SfmCAD to many different types of objects, such as CAD parts, ShapeNet objects, and tree shapes. Extensive comparisons show that SfmCAD produces compact and faithful 3D reconstructions with superior quality compared to alternatives. The code is released at https://***/BunnySoCrazy/SfmCAD.
暂无评论