scholarly journals Multi Camera System Analysis for Autonomous Navigation using End-to-End Deep Learning

2019 ◽  
Author(s):  
José A. Diaz Amado ◽  
Jean Amaro ◽  
Iago P. Gomes ◽  
Denis Wolf ◽  
F. S. Osorio
Keyword(s):  
Deep Learning ◽  
Autonomous Navigation ◽  
System Analysis ◽  
Autonomous Vehicle ◽  
Small Scale ◽  
Camera System ◽  
Depth Images ◽  
End To End ◽  
Autonomous Vehicle Navigation ◽  
The Impact

This work aims to present an autonomous vehicle navigation system, based on an End-to-End Deep Learning approach, and to study the impact of different image input configurations to the system performance. The proposed methodology in this work was to adoptand test different configurations of RGB and Depth images captured from a Kinect device. We adopted a multi-camera system, composed by 3 cameras, with different RGB and/or Depth input configurations. Two main systems were developed in order to study and validade de different input configurations: the first one based on a realistic simulator and the second one based on a mini-car (small scale vehicle). Starting with the simulations, it was possible to choose the best camera/input configuration, then we validated that using the real vehicle (mini-car) with real sensors/cameras. The experimental results demonstrated that a multi-camera solution, based on 3 cameras, allow us to obtain better autonomous navigation control results in a End-to-End Deep Learning based approch, with a very small final error when using the proposed camera configurations.

End-to-End Deep Learning Applied in Autonomous Navigation using Multi-Cameras System with RGB and Depth Images

2019 ◽  
Author(s):  
Jose A. Diaz Amado ◽  
Iago Pacheco Gomes ◽  
Jean Amaro ◽  
Denis Fernando Wolf ◽  
Fernando S. Osorio
Keyword(s):  
Deep Learning ◽  
Autonomous Navigation ◽  
Depth Images ◽  
End To End

An Experimental Analysis of the Effects of Different Hardware Setups on Stereo Camera Systems

2021 ◽  
Vol 15 (03) ◽  
pp. 337-357
Author(s):  
Alexander Julian Golkowski ◽  
Marcus Handte ◽  
Peter Roch ◽  
Pedro J. Marrón
Keyword(s):  
Experimental Analysis ◽  
Autonomous Navigation ◽  
Indoor Environments ◽  
Camera System ◽  
Stereo Camera ◽  
Camera Systems ◽  
Sensing Technology ◽  
Stereo Cameras ◽  
The Impact ◽  
General Guidance

For many application areas such as autonomous navigation, the ability to accurately perceive the environment is essential. For this purpose, a wide variety of well-researched sensor systems are available that can be used to detect obstacles or navigation targets. Stereo cameras have emerged as a very versatile sensing technology in this regard due to their low hardware cost and high fidelity. Consequently, much work has been done to integrate them into mobile robots. However, the existing literature focuses on presenting the concepts and algorithms used to implement the desired robot functions on top of a given camera setup. As a result, the rationale and impact of choosing this camera setup are usually neither discussed nor described. Thus, when designing the stereo camera system for a mobile robot, there is not much general guidance beyond isolated setups that worked for a specific robot. To close the gap, this paper studies the impact of the physical setup of a stereo camera system in indoor environments. To do this, we present the results of an experimental analysis in which we use a given software setup to estimate the distance to an object while systematically changing the camera setup. Thereby, we vary the three main parameters of the physical camera setup, namely the angle and distance between the cameras as well as the field of view and a rather soft parameter, the resolution. Based on the results, we derive several guidelines on how to choose the parameters for an application.

scholarly journals End-to-End Deep Learning Fusion of Fingerprint and Electrocardiogram Signals for Presentation Attack Detection

Sensors ◽  
2020 ◽  
Vol 20 (7) ◽  
pp. 2085 ◽  
Author(s):  
Rami M. Jomaa ◽  
Hassan Mathkour ◽  
Yakoub Bazi ◽  
Md Saiful Islam
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Attack Detection ◽  
Ecg Signal ◽  
Ecg Signals ◽  
Biometric Systems ◽  
Fingerprint Biometrics ◽  
End To End ◽  
The Impact

Although fingerprint-based systems are the commonly used biometric systems, they suffer from a critical vulnerability to a presentation attack (PA). Therefore, several approaches based on a fingerprint biometrics have been developed to increase the robustness against a PA. We propose an alternative approach based on the combination of fingerprint and electrocardiogram (ECG) signals. An ECG signal has advantageous characteristics that prevent the replication. Combining a fingerprint with an ECG signal is a potentially interesting solution to reduce the impact of PAs in biometric systems. We also propose a novel end-to-end deep learning-based fusion neural architecture between a fingerprint and an ECG signal to improve PA detection in fingerprint biometrics. Our model uses state-of-the-art EfficientNets for generating a fingerprint feature representation. For the ECG, we investigate three different architectures based on fully-connected layers (FC), a 1D-convolutional neural network (1D-CNN), and a 2D-convolutional neural network (2D-CNN). The 2D-CNN converts the ECG signals into an image and uses inverted Mobilenet-v2 layers for feature generation. We evaluated the method on a multimodal dataset, that is, a customized fusion of the LivDet 2015 fingerprint dataset and ECG data from real subjects. Experimental results reveal that this architecture yields a better average classification accuracy compared to a single fingerprint modality.

scholarly journals UAV-g 2019: Unmanned Aerial Vehicles in Geomatics

Drones ◽  
2019 ◽  
Vol 3 (3) ◽  
pp. 74 ◽  
Author(s):  
Nex
Keyword(s):  
Remote Sensing ◽  
Deep Learning ◽  
Unmanned Aerial Vehicles ◽  
Autonomous Navigation ◽  
Aerial Vehicles ◽  
Surrounding Environment ◽  
Poster Sessions ◽  
Aerial Vehicle ◽  
The University ◽  
The Impact

Unmanned aerial vehicle in geomatics (UAV-g) is a well-established scientific event dedicated to UAVs in geomatics and remote sensing. In the different editions of the journal, new scientific challenges have increased their synergy with adjacent domains, such as robotics and computer vision, thereby increasing the impact of this conference. The 2019 edition has been hosted by the University of Twente (The Netherlands) and has attracted about 300 participants for the full three-day program. Researchers from 36 different countries (from all continents) have presented 89 accepted papers in 17 oral and 2 poster sessions. The presented papers covered multi-disciplinary topics, such as photogrammetry, natural resources monitoring, autonomous navigation, and deep learning. All these contributions have in common the use of UAV platforms for the innovative acquisition and processing of the acquired data and information extracted from the surrounding environment.

scholarly journals Training confounder-free deep learning models for medical applications

2020 ◽  
Vol 11 (1) ◽  
Author(s):  
Qingyu Zhao ◽  
Ehsan Adeli ◽  
Kilian M. Pohl
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Bone Age ◽  
Magnetic Resonance Images ◽  
Data Sets ◽  
Large Set ◽  
Learning Models ◽  
End To End ◽  
The Impact ◽  
The Relationship

AbstractThe presence of confounding effects (or biases) is one of the most critical challenges in using deep learning to advance discovery in medical imaging studies. Confounders affect the relationship between input data (e.g., brain MRIs) and output variables (e.g., diagnosis). Improper modeling of those relationships often results in spurious and biased associations. Traditional machine learning and statistical models minimize the impact of confounders by, for example, matching data sets, stratifying data, or residualizing imaging measurements. Alternative strategies are needed for state-of-the-art deep learning models that use end-to-end training to automatically extract informative features from large set of images. In this article, we introduce an end-to-end approach for deriving features invariant to confounding factors while accounting for intrinsic correlations between the confounder(s) and prediction outcome. The method does so by exploiting concepts from traditional statistical methods and recent fair machine learning schemes. We evaluate the method on predicting the diagnosis of HIV solely from Magnetic Resonance Images (MRIs), identifying morphological sex differences in adolescence from those of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA), and determining the bone age from X-ray images of children. The results show that our method can accurately predict while reducing biases associated with confounders. The code is available at https://github.com/qingyuzhao/br-net.

scholarly journals Double Deep Q-Learning and Faster R-CNN-Based Autonomous Vehicle Navigation and Obstacle Avoidance in Dynamic Environment

Sensors ◽  
2021 ◽  
Vol 21 (4) ◽  
pp. 1468
Author(s):  
Razin Bin Issa ◽  
Modhumonty Das ◽  
Md. Saferi Rahman ◽  
Monika Barua ◽  
Md. Khalilur Rhaman ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Autonomous Navigation ◽  
Autonomous Vehicle ◽  
Dynamic Environment ◽  
Vehicle Navigation ◽  
Autonomous Agent ◽  
Learning Approach ◽  
Q Learning ◽  
Land Vehicles ◽  
Autonomous Vehicle Navigation

Autonomous vehicle navigation in an unknown dynamic environment is crucial for both supervised- and Reinforcement Learning-based autonomous maneuvering. The cooperative fusion of these two learning approaches has the potential to be an effective mechanism to tackle indefinite environmental dynamics. Most of the state-of-the-art autonomous vehicle navigation systems are trained on a specific mapped model with familiar environmental dynamics. However, this research focuses on the cooperative fusion of supervised and Reinforcement Learning technologies for autonomous navigation of land vehicles in a dynamic and unknown environment. The Faster R-CNN, a supervised learning approach, identifies the ambient environmental obstacles for untroubled maneuver of the autonomous vehicle. Whereas, the training policies of Double Deep Q-Learning, a Reinforcement Learning approach, enable the autonomous agent to learn effective navigation decisions form the dynamic environment. The proposed model is primarily tested in a gaming environment similar to the real-world. It exhibits the overall efficiency and effectiveness in the maneuver of autonomous land vehicles.

scholarly journals Ultra-Wideband Positioning Sensor with Application to an Autonomous Ultraviolet-C Disinfection Vehicle

Sensors ◽  
2021 ◽  
Vol 21 (15) ◽  
pp. 5223
Author(s):  
Shih-Ping Huang ◽  
Jin-Feng Neo ◽  
Yu-Yao Chen ◽  
Chien-Bang Chen ◽  
Ting-Wei Wu ◽  
...  
Keyword(s):  
Autonomous Navigation ◽  
Autonomous Vehicle ◽  
Ultra Wideband ◽  
Average Error ◽  
Indoor Environments ◽  
Optimal Position ◽  
Reproductive Ability ◽  
High Efficient ◽  
Ultraviolet C ◽  
Autonomous Vehicle Navigation

Due to the COVID-19 virus being highly transmittable, frequently cleaning and disinfecting facilities is common guidance in public places. However, the more often the environment is cleaned, the higher the risk of cleaning staff getting infected. Therefore, strong demand for sanitizing areas in automatic modes is undoubtedly expected. In this paper, an autonomous disinfection vehicle with an Ultraviolet-C (UVC) lamp is designed and implemented using an ultra-wideband (UWB) positioning sensor. The UVC dose for 90% inactivation of the reproductive ability of COVID-19 is 41.7 J/m2, which a 40 W UVC lamp can achieve within a 1.6 m distance for an exposure time of 30 s. With this UVC lamp, the disinfection vehicle can effectively sterilize in various scenarios. In addition, the high-accuracy UWB positioning system, with the time difference of arrival (TDOA) algorithm, is also studied for autonomous vehicle navigation in indoor environments. The number of UWB tags that use a synchronization protocol between UWB anchors can be unlimited. Moreover, this proposed Gradient Descent (GD), which uses Taylor method, is a high-efficient algorithm for finding the optimal position for real-time computation due to its low error and short calculating time. The generalized traversal path planning procedure, with the edge searching method, is presented to improve the efficiency of autonomous navigation. The average error of the practical navigation demonstrated in the meeting room is 0.10 m. The scalability of the designed system to different application scenarios is also discussed and experimentally demonstrated. Hence, the usefulness of the proposed UWB sensor applied to UVC disinfection vehicles to prevent COVID-19 infection is verified by employing it to sterilize indoor environments without human operation.

Sign in / Sign up

Export Citation Format

Share Document