Tactile sensors are the last frontier to robots that can handle everyday objects and interact with humans through contact. To be effective, such have sense geometry of touched surfaces objects, as well any other relevant information for their tasks, forces, vibrations, temperature, allow them safely securely within an environment. Given capability easily capture interpret tactile data, one promising direction in order produce enhanced robotic is explore imitate human sensing capabilities. In...

Effective human-like manipulation in robots depends on their capacity to recognize and identify textures different environments. In unpredictable environments, with tactile sensors will have through touch-related features. To advance research texture classification, a comprehensive dataset capturing the physical interactions between tactile-enabled robotic probe various is necessary. As result, we are driven create from signals collected by bioinspired multimodal sensing module, as...

Robots need to adapt the complexities of acting in unstructured environments. Robotic systems more awareness environment uncertainty and variability. Although cameras have been predominantly used robotic tasks, limitations that come with them, such as occlusion, visibility, lack information, diverted some focus tactile sensing. Extensive datasets physical interactions between tactile-enabled robots are required investigate develop methods for performing manipulation object exploration tasks....

This work proposes a map alignment approach based on simple modification of the Iterative Closest Point algorithm to use point confidences metrics usually available in feature-based visual SLAM pipelines. In context hierarchical composed mini-maps, aligning these mini-maps is an important task allow metric information be related between them. research enumerates three possible that could used for representing landmark confidence, and investigate potential using improve ICP algorithm. The...

Robots are expected to recognize the properties of objects in order handle them safely and efficiently a variety applications, such as health elder care, manufacturing, or high-risk environments. This paper explores issue surface characterization by monitoring signals acquired novel bio-inspired tactile probe contact with ridged surfaces. The module comprises nine Degree Freedom Microelectromechanical Magnetic, Angular Rate, Gravity system (9-DOF MEMS MARG) deep pressure sensor embedded...

This paper proposes a set of evolving Takagi-Sugeno-Kang (TSK) fuzzy models the nonlinear dynamic mechanisms occurring in myoelectric-based control prosthetic hand fingers. The rule bases and parameters TSK are evolved by an online identification algorithm. experimental results prove performance good output responses root mean square error values. A comparison with two recurrent neural network architectures is included.

A new generation of humanoid robots is emerging to work together with, or even replace, human operators performing complex dextrous manipulation operations in a variety applications such as health and elder care, hazardous high-risk environments, telemedicine, manufacturing. To meet the challenging operational requirements applications, this should not only look humans, but also behave like them, being able sense perceive external world perform tasks humans do. Touch sensing perception...

Underactuated hands are useful tools for robotic in-hand manipulation tasks due to their capability seamlessly adapt unknown objects. To enable robots using such achieve and maintain stable grasping conditions even under external disturbances while keeping track of an object's state requires learning object-tactile sensing data relationships. The human somatosensory system combines visual tactile information in "What Where" subsystem high levels skills. present paper proposes approach...

The paper discusses signal processing and mechatronics solutions for tactile object recognition using robotic hands. two important aspects analyzed are the of objects from displacement profiles obtained by force sensing transducers textures a rubbing motion executed finger equipped with dynamic fingertip. Neural network used as an intelligent approach to recognize symbols, such embossed numbers letters, well different texture profiles. conception issues, challenges experimental results...

This paper discusses a novel nonlinear autoregressive network with exogenous inputs (NARX) neural (NN) model of the dynamic mechanisms occurring in myoelectric-based control prosthetic hand fingers. The experimental results demonstrate good performance NN architecture terms output response and mean squared error. comparison also shows that proposed NARX recurrent outperforms linear same training algorithm.

We present the open-source design and fabrication of a compliant multimodal tactile sensing module. The module presented here enables robotic end-effectors to sense contact properties like pressure vibration estimate quaternion that represents deformation due contact. elaborated module's structure process use only 3D printed molds vacuum chamber, making it accessible broad range roboticists. also allows production batches up five modules easy deployment hands. repository associated contains...

This paper presents a novel vision-based hand gesture recognition (HGR) and training system for human-like robot hand. We implemented trained multiclass-SVM classifier N-Dimensional DTW (ND-DTW) static posture dynamic recognition. Training features were extracted from the raw gestures depth data captured by Leap Motion Controller. The experimental results show that multiclass SVM method has an average 98.25% rates shortest run time when compared to k-NN ANBC. For gestures, ND-DTW displays...

Fibrosis is a significant indication of chronic liver diseases often due to hepatitis C Virus. It becoming global concern as result the rapid increase in number HCV infected patients, high cost and flaws associated with assessment process fibrosis. This study aims determine features that significantly contribute identification stages fibrosis generate rules assist physicians during treatment patients clinically non-invasive approach. Also, performance different Multi-layered Perceptron...

The findings of this paper enhance the knowledge machine learning in tactile texture recognition. Synthetic perception textures was achieved using a recently developed miniaturized multi-modal sensor. Thirteen commonly used were dynamically explored and classified supervised classifiers. best result by Extra Trees classifier with an average accuracy score 95%. Other classifiers also had high-performance results 92% (Random Forest), 91% (Support-Vector Machines), 88% (Multilayer Perceptron)....

In this work we investigate the exploration of an environment with multiple vehicles using a strategy based on occupancy grids and technique simultaneous localization mapping (SLAM). The uses concepts costs utility from frontier cells. Besides, used SLAM method is FastSLAM algorithm landmarks extracted visual sensors common map characteristics. Both activities are coordinated by central agent. results obtained simulation show that when two can communicate agent building characteristics, task...

A new approach to the in-hand manipulation problem, using fuzzy controlled teleoperation and a recently proposed multimodal tactile sensor are presented in this article. Real-time grasping experiments were performed by gripper, with compliant mounted on gripper. As for robot dexterity, teleoperated thumb object rotation while gripper maintained stable grasping. The data from consistent and, thus, provided substantial tool control algorithms.

This paper presents a new design of tactile-enabled open-source tendon driven robotic hand. The upgraded hand, so-called `Ring Ada,' comprises 3D-printed `rings' used to better control the compliant fingers. Experiments were conducted enhance postural synergy-based teleoperated grasping using multi-finger hand and haptic-feedback data glove for remote human operators.

Social media platforms have become the most prominent medium for spreading hate speech, primarily through hateful textual content. An extensive dataset containing emoticons, emojis, hashtags, slang, and contractions is required to detect speech on social based current trends. Therefore, our curated from various sources like Kaggle, GitHub, other websites. This contains sentences in English confined into two classes, one representing content non-hateful It has 451,709 total. 371,452 of these...

Reproducing human-like dexterous manipulation in robots requires identifying objects and textures. In unstructured settings, equipped with tactile sensors may detect textures by using touch-related characteristics. An extensive dataset of the physical interaction between a tactile-enable robotic probe is required to investigate develop methods for categorizing Therefore, this motivates us compose from signals bioinspired multimodal sensing module while brings dynamically contact 12 under...

Robotic hands able to conduct precision in-hand manipulation have become imperative for modern robotic systems. Compliant finger joints provide compliant grasping but the post-grasp orientation of an object varies. Under-actuated, that are capable and reorienting objects one step closer human-like end-effectors. This paper presents a hand prototype with two under-actuated fingers embedded tactile feedback fuzzy system obtain stable grasp, integrated gripper. One fully actuated thumb 3...

The paper presents a heart rate detection system using novel miniaturized multimodal tactile sensor that emulates the four main mechanoreceptors - Meissner's corpuscles, Merkel disks, Ruffini, and Pacinian corpuscles of human skin. A wrist monitor was used to validate performance new proposed system. experiment consisted placing sensors on user's carotid artery. In 80% cases, control measured similar values, with difference up 2 bpm. other results had external interference due misplacement