This paper provides a methodology for dynamic gesture recognition for enhanced Human-Robot Collaboration in search and rescue missions. This research was conducted at DFKI in collaboration with THW, Germany, where an ...
详细信息
ISBN:
(纸本)9798331510961;9798331510954
This paper provides a methodology for dynamic gesture recognition for enhanced Human-Robot Collaboration in search and rescue missions. This research was conducted at DFKI in collaboration with THW, Germany, where an autonomous robot is built for the assistance of emergency workers during disaster scenarios for transporting the materials. This assistance robot will be highly functional in risk-prone areas where the safety of workers is in danger. A set of standard gestures were predefined and those will be used as commands from humans to the robot. Due to the criticality of the area of application of this assistance robot, the identification and interpretation of the gestures must be in real-time and extremely accurate. In this paper, we compare different methods and train different pipelines capable of performing dynamic gesture recognition. The final model pipeline can understand the difference between static as well as dynamic gestures and recognize both of them accurately. The key challenge to attaining real-time gesture recognition is the models's capability to give rapid and highly accurate responses.
This paper presents an adaptive online learning framework for systems with uncertain parameters to ensure safety-critical control in non-stationary environments. Our approach consists of two phases. The initial phase ...
详细信息
ISBN:
(纸本)9798350384581;9798350384574
This paper presents an adaptive online learning framework for systems with uncertain parameters to ensure safety-critical control in non-stationary environments. Our approach consists of two phases. The initial phase is centered on a novel sparse Gaussian process (GP) framework. We first integrate a forgetting factor to refine a variational sparse GP algorithm, thus enhancing its adaptability. Subsequently, the hyperparameters of the Gaussian model are trained with a specially compound kernel, and the Gaussian model's online inferential capability and computational efficiency are strengthened by updating a solitary inducing point derived from newly samples, in conjunction with the learned hyperparameters. In the second phase, we propose a safety filter based on high order control barrier functions (HOCBFs), synergized with the previously trained learning model. By leveraging the compound kernel from the first phase, we effectively address the inherent limitations of GPs in handling high-dimensional problems for real-time applications. The derived controller ensures a rigorous lower bound on the probability of satisfying the safety specification. Finally, the efficacy of our proposed algorithm is demonstrated through real-time obstacle avoidance experiments executed using both simulation platform and a real-world 7-DOF robot.
Scoliosis diagnosis and assessment depend largely on the measurement of the Cobb angle in spine X-ray images. With the emergence of deep learning techniques that employ landmark detection, tilt prediction, and spine s...
详细信息
ISBN:
(纸本)9798350384581;9798350384574
Scoliosis diagnosis and assessment depend largely on the measurement of the Cobb angle in spine X-ray images. With the emergence of deep learning techniques that employ landmark detection, tilt prediction, and spine segmentation, automated Cobb angle measurement has become increasingly popular. However, these methods encounter difficulties such as high noise sensitivity, intricate computational procedures, and exclusive reliance on a single type of morphological information. In this paper, we introduce the Multiple Morphology-Aware Network (MMA-Net), a novel framework that improves Cobb angle measurement accuracy by integrating multiple spine morphology as attention information. In the MMA-Net, we first feed spine X-ray images into the segmentation network to produce multiple morphological information (spine region, centerline, and boundary) and then concatenate the original X-ray image with the resulting segmentation maps as input for the regression module to perform precise Cobb angle measurement. Furthermore, we devise joint loss functions for our segmentation and regression network training, respectively. We evaluate our method on the AASCE challenge dataset and achieve superior performance with the SMAPE of 7.28% and the MAE of 3.18 degrees, indicating a strong competitiveness compared to other outstanding methods. Consequently, we can offer clinicians automated, efficient, and reliable Cobb angle measurement.
In this paper we consider the problem of allowing a humanoid robot that is subject to a persistent disturbance, in the form of a payload-carrying task, to follow given planned footsteps. To solve this problem, we comb...
详细信息
ISBN:
(纸本)9798350323658
In this paper we consider the problem of allowing a humanoid robot that is subject to a persistent disturbance, in the form of a payload-carrying task, to follow given planned footsteps. To solve this problem, we combine an online non-linear centroidal Model Predictive Controller - MPC with a contact stable force parametrization. The cost function of the MPC is augmented with terms handling the disturbance and regularizing the parameter. The performance of the resulting controller is validated both in simulations and on the humanoid robot iCub. Finally, the effect of using the parametrization on the computational time of the controller is briefly studied.
Dubins coverage has been extensively researched to address the coverage path planning (CPP) problem of a known environment for the curvature-constrained robot. However, its fixed-speed assumption prevents the robot fr...
详细信息
ISBN:
(纸本)9798350323658
Dubins coverage has been extensively researched to address the coverage path planning (CPP) problem of a known environment for the curvature-constrained robot. However, its fixed-speed assumption prevents the robot from accelerating to reduce the time and limits its flexibility to avoid obstacles. Therefore, this paper presents a collision-free CPP approach (CFC) for the obstacle-constrained environment, which enhances time efficiency by constructing the variable-speed Dubins paths and ensures robot safety by building a risk potential surface for representing the possibility of collision. Furthermore, CFC models the CPP problem as an asymmetric traveling salesman problem (ATSP) and utilizes a graph pruning strategy to reduce the computational cost. Comparison tests with other Dubins coverage methods demonstrate that CFC provides shorter coverage times and better runtimes than the other Dubins coverage methods while preventing collision risk between the robot and obstacles. Physical experiments in a laboratory setting demonstrate the applicability of CFC to the physical robot.
This study presents a novel approach to Human-Robot Interaction (HRI) by employing gesture recognition to control robotic systems, such as bionic arms or manipulators, through accelerometer-based data collection. Usin...
详细信息
People with diabetes, clinicians and researchers alike need appropriate metrics to manage diabetes, assess the quality of blood glucose control, and guide the development of new therapies and control methods. In this ...
详细信息
ISBN:
(纸本)9798350398823
People with diabetes, clinicians and researchers alike need appropriate metrics to manage diabetes, assess the quality of blood glucose control, and guide the development of new therapies and control methods. In this paper, we review some commonly used metrics with their advantages and disadvantages. We analyze the glucose data of a CGM user to compare the results of different metrics and determine which of these gives the most complete picture of glycemic control in silico trials of artificial pancreas systems.
Transparent slide detection, pick and place automation are obtained in order to boost speed, accuracy and smooth workflow between slide coverslipping and digital scanning operation in digital pathology. Fully automati...
详细信息
ISBN:
(纸本)9798350398823
Transparent slide detection, pick and place automation are obtained in order to boost speed, accuracy and smooth workflow between slide coverslipping and digital scanning operation in digital pathology. Fully automatic, precise, robust solution has been developed by robotic arm using computer vision and image processing techniques. Magazine (slide holder) and gripper design, camera selection and mapping to robot base coordinate system are all guided. Python based software algorithms for slide detection and 3D position determination applied in dynamically changeable environment are also explained step by step. Complete automatic process control method is implemented and tested for different types of slide.
Cloud robotics is transforming industrial automation by combining cloud computing with robotic systems. This paper reviews recent advancements in cloud robotics and their industrial applications, highlighting the syne...
详细信息
作者：
Balint, KrisztianObudai Univ
Keleti Karoly Fac Business & Management Inst Enterprise Dev & Infocommun Budapest Hungary
The latest, modern security camera systems record numerous data at once. With the utilization of artificial intelligence, these systems can even compose an online attendance register of students present during the lec...
详细信息
ISBN:
(纸本)9798350398823
The latest, modern security camera systems record numerous data at once. With the utilization of artificial intelligence, these systems can even compose an online attendance register of students present during the lectures. Data is primarily recorded on the hard disk of the NVR (Network Video Recorder), and in the long term, it is recommended to save the data in the blockchain. The purpose of the research is to demonstrate how university security cameras can be securely connected to the blockchain. This would be important for universities as this is sensitive student data that needs to be protected from unauthorized access. In my research, as part of the practical implementation, I therefore also use encryption methods and data fragmentation, which are saved at the nodes of the blockchain. Thus, even a DDoS (Distributed Denial of Service) type attack may be easily repelled, as data is not concentrated on a single, central server. To further increase security, it is useful to constitute a blockchain capable of its own data storage at the faculty itself, rather than renting data storage space, so we, ourselves may regulate the conditions of operation, and the policy of data protection. As a practical part of my research, therefore, I created a blockchain called UEDSC (Universities Data Storage Chain) where I saved the student's data.
暂无评论