A Sliding-Mode Control Algorithm to Enhance In-Hand Motion Capabilities

This paper reports a method for regulating the internal forces during in hand manipulation of an unknown shaped object with soft robotic fingers. The internal forces ensure that the object does not move between the robotic fingers, thus improving the grip. It is shown that if soft fingers show bounded conformity and the finger-object interface have bounded relative slip, then the relative angular velocity between the object and the fingertip coordinate frame in contact is bounded. Detailed derivation of the proof is presented. The proof is used to define a new metric of relative slip. The metric is used to design a sliding mode control algorithm that results in an efficient grip which is robust towards uncertainty in object shape. The robotic fingers are assumed to be under virtual rigidity constraint, that is, the distance between the fingertips do not change. The control algorithm is attractive as it skirts the requirement of information of the shape of the object or to solve optimization problems. The grip with the robust control algorithm is shown to be finite-time stable through Lyapunov's method. The methodology is demonstrated using simulations.

Deciphering 3D Human Sphygmopalpation Pulse Patterns using “X-ray” Images Acquired from Tactile Robotic Fingers

ABSTRACTSphygmopalpation at specific locations of human wrists has been used as a medical measurement technique in China since the Han Dynasty (202 BC - 220 AD); it is now generally accepted that traditional Chinese medicine (TCM) doctors are able to decipher 28 types of basic pulse patterns using their fingertips. This TCM technique of examining individual arterial pulses by palpation has undergone an upsurge recently in popularity as a low-cost and non-invasive diagnostic technique for monitoring patient health status. We have developed a pulse sensing platform for studying and digitalizing arterial pulse patterns via a TCM approach. This platform consists of a robotic hand with three fingers for pulse measurement and an artificial neural network (ANN) together with pulse signal preprocessing for pulse pattern recognition. The platforms previously reported by other research groups or marketed commercially exhibit one or more of the following imperfections: a single channel for data acquisition, low sensitivity and rigid sensors, lack of control of the applied pressure, and in many reported works, lack of an intelligent data analysis system. The platform presented here features up to three-dimensional (3D) tactile sensing channels for recording data and uses highly sensitive capacitive MEMS (microelectromechanical systems) flexible sensing arrays, pressure-feedback-controlled robotic fingers, and machine learning algorithms. We also proposed a methodology of obtaining “X-ray” image of pulse information constructed based on the sensing data from 3 locations and 3 applied pressures (i.e., mimicking TCM doctors), which contains all arterial pulse information in both spatial and temporal spans, and which could be used as an input to a deep learning algorithm. By applying our developed platform and algorithms, 3 types of consistent pulse patterns, i.e., “Hua” , “Xi” , and “Chen” , as described by TCM doctors”, could be identified in a selected group of 3 subjects who were diagnosed by TCM practitioners. We have shown the classification rates is 98.7% in training process and 84.2% in testing result for these 3 basic pulse patterns. The high classification rate of the developed platform could lead to further development of a high-level artificial intelligence system incorporating knowledge from TCM – the robotics finger system could become a standard clinical equipment for digitalizing and visualizing human arterial pulses

Deformation Modeling of Compliant Robotic Fingers Grasping Soft Objects

In this paper, a model is shown to predict the simultaneous deformations occurring when compliant robotic fingers are grasping soft objects. This model aims at providing an accurate estimation of the penetration, internal forces, and deformed shapes of both these fingers and the objects. A particular emphasis is placed on the case when the finger is underactuated but the methodology discussed in this paper is general. Usually in the literature, underactuated fingers are modeled and designed considering their grasps of rigid object because of the complexity associated with deforming objects. This limitation severely hinders the usability of underactuated grippers and either restricts them to a narrow range of applications or requires extensive experimental testing. Furthermore, classical models of underactuated fingers in contact with objects are typically applicable with a maximum of one contact per phalanx only. The model proposed in this paper demonstrates a simple algorithm to compute a virtual subdivision of the phalanges which can be used to estimate the contact pressure arising when there are contacts at many locations simultaneously. This model also proposes a computationally efficient approximation of isotropic soft objects. Numerical simulations of the proposed model are compared here with dynamic simulations, finite element analyses, and experimental measurements which all shows its effectiveness and accuracy. Finally, the extension of the model to other types of underactuated fingers, standard grippers, and fully actuated robotic fingers as well as potential applications is discussed and illustrated.

A Sliding-Mode Control Algorithm to Enhance In-Hand Motion Capabilities

This paper reports a method for regulating the internal forces during in hand manipulation of an unknown shaped object with soft robotic fingers. It is known that for the case of multifingered manipulation, a part of the forces applied by the fingers result in the motion of the object, whereas the other part is considered to be an internal force. The internal forces do not result in the motion of the object but are used to improve the grip on the object. For an object with unknown shape, the internal forces are regulated to ensure that the object does not slip off during manipulation. It is shown that if soft fingers show bounded conformity and the finger-object interface does not have relative slip (or a bounded slip), then the relative angular velocity between the object and the fingertip frame in contact is bounded. The proof is used to define of a new metric of relative slip. The metric is used to design a sliding mode control algorithm. The robotic fingers are assumed to be under virtual rigidity constraint, that is, the distance between the fingers do not change. The control algorithm is attractive as it skirts requirement of information of the shape of the object or to solve optimization problems. The control algorithm developed controls the internal forces and does not require the knowledge of the shape of the object. The methodology is simulated for the case of one spherical object and one conical object.

Grasp Posture Control of Wearable Extra Robotic Fingers with Flex Sensors Based on Neural Network

This study proposes a data-driven control method of extra robotic fingers to assist a user in bimanual object manipulation that requires two hands. The robotic system comprises two main parts, i.e., robotic thumb (RT) and robotic fingers (RF). The RT is attached next to the user’s thumb, while the RF is located next to the user’s little finger. The grasp postures of the RT and RF are driven by bending angle inputs of flex sensors, attached to the thumb and other fingers of the user. A modified glove sensor is developed by attaching three flex sensors to the thumb, index, and middle fingers of a wearer. Various hand gestures are then mapped using a neural network. The input data of the robotic system are the bending angles of thumb and index, read by flex sensors, and the outputs are commanded servo angles for the RF and RT. The third flex sensor is attached to the middle finger to hold the extra robotic finger’s posture. Two force-sensitive resistors (FSRs) are attached to the RF and RT for the haptic feedback when the robot is worn to take and grasp a fragile object, such as an egg. The trained neural network is embedded into the wearable extra robotic fingers to control the robotic motion and assist the human fingers in bimanual object manipulation tasks. The developed extra fingers are tested for their capacity to assist the human fingers and perform 10 different bimanual tasks, such as holding a large object, lifting and operate an eight-inch tablet, and lifting a bottle, and opening a bottle cap at the same time.