Project Logs

Collapse

• Results: Wearable Muscle-machine Interface

  Anany Dwivedi • 10/05/2020 at 07:47 • 0 comments

  To control the proposed motorized exoskeleton glove, the wearable muscle-machine interface was used. It combines Electromyography (EMG) and Forcemyography (FMG) sensors to decode the user’s intentions and discriminates between different grasp types. In order to develop the intention decoding models Linear Discriminant Analysis (LDA) based learning scheme was selected after considering the trade off between accuracy of classifying the grasps and the time taken to make the predictions. The other techniques that were considered were, Random Forests (RF) and Support Vector Machines (SVM). The examined RF based models were trained with 100 trees and max depth of each tree as 10. For the SVM classifier we used a non-linear RBF kernel. Along with the type of classifier, the performance for the two types of data (EMG and FMG) was also evaluated. More precisely, the training data was divided into 3 different sets. In Set 1, only the data from the EMG sensors was used for training the classification model. In Set 2, only the data from the FMG sensors was used for training and in Set 3, data from both EMG and FMG sensors was used for training. The classification performance over the 5-fold cross validation method for the three different classifiers and the different sets of data is presented in Table I. The execution time to predict the grasp types for 10,000 data points (samples) for the three examined classifiers is shown in Table II.
  

• Laminar Jamming Structures for Variable Stiffness Fingers

  Lucas Gerez • 10/05/2020 at 07:35 • 0 comments

  The soft, laminar jamming structure offers multiple stiffness by applying a pressure gradient to the system. The figure below shows a 100 g weight being used to demonstrate the structure’s ability to resist deflecting loads when jammed (high stiffness) and to freely deform when unjammed (low stiffness). The structure consists of a soft pouch with several thin layers inside it. When the vacuum is applied inside the pouch, the high friction between layers stiffens the entire structure.  Each jamming structure contains seven thin layers encased in the silicone vacuum pouch. Each layer is made out of a urethane rubber matrix (Smooth-On Vytaflex 30) and carbon fiber sheets. The composite layers retain the high flexibility of the rubber (promoting passive extension of the fingers) and low deformation in the axial direction of the carbon fiber (increasing the bending resistance when the structure is under vacuum). 

  The elastic properties of the jamming layers and the soft pouch promote the passive extension of the fingers, keeping the human hand in a natural, zero-effort pose, but can also be used to hold the hand in certain configurations when jammed. 
  

• Smartphone Based Control of Hybrid, Motorized Exoskeleton Glove

  Lucas Gerez • 10/05/2020 at 07:18 • 0 comments

  The hybrid, motorized exoskeleton glove uses an intuitive smartphone based interface,  which allows the user selects the mode desired to control the exoskeleton glove. The user can combine the motions (e.g., full grasp with abducted fingers or tripod grasp with the extra thumb inflated and the jamming structures activated). The flex sensor can be used in order to trigger the desired function when a set bending angle is reached. The desired function is selected from a predefined set stored within the smartphone app, like tripod grasp, full grasp, or jamming structures triggering, as seen in the figures below. The information is transmitted to a microcontroller through Bluetooth communication. Then, the microcontroller activates the chosen actuators that are connected to the glove. Each air supply is connected to a solenoid valve that is used to release the air pressure of the system when the glove returns to its original position. The entire system is powered by a 12V power supply. A battery can be added to the control unit to make the device portable and wearable. The amount of force applied by each finger is determined through a current control of each motor and can be adjusted according to the user’s needs. 

• Grasp Decoding with a Wearable Muscle-machine Interface

  Lucas Gerez • 10/04/2020 at 18:22 • 0 comments

  The wearable sleeve muscle-machine interface developed for the control of the proposed glove system can be employed to decode various gestures/grasps, like pinch grasps, power grasps, tripod grasps, key grasps, spherical grasp and rest state. With these five different types of grasp it is possible to perform most of Activities of Daily Living (ADLs). Fig. 1 shows the five gestures. An example of the acquired signal for the power and pinch grasp can be seen in Fig. 2.

  For user intention classification the machine learning based models can be developed which use the EMG and FMG data from eight different muscle sites collected using the wearable sensorized sleeve. At a particular instance in time, the input data vector for training the learning model can be represented as:

  where xt1, xt2, xt3 are values for the EMG sensors E1, E2, and E3 at a time instance ‘t’. While represent xt4,  xt5,  xt6, xt7, xt8 values of the FMG sensors F1, F2, F3, F4, and F5 at time ‘t’. The desired output of the learned model at time ‘t’ can be represented as:

  where Ht= SP corresponds to the spherical grasp, Ht= PO corresponds to the power grasp, Ht= PI corresponds to the pinch grasp, Ht= TP corresponds to the tripod grasp, Ht= K corresponds to the key grasp, while Ht= R corresponds to the rest state of the hand at time ‘t’. For each of the intended grasp motion Ht, a predefined Mt R4 that correspond to the motor state for each of the grasp strategies. For each of the grasp types a specific Mt is triggered for the proposed glove to execute the corresponding grasping motion. For a robust classification outcome, we use the Majority Vote Criterion (MVC). To do this, a sliding window, of size W = 10 was applied on the data while performing predictions. The MVC classifies all the samples in the window as the class that received the maximum number of votes in that window.

  The code for the wearable muscle-machine interface (trainer, predictions, and operation) is available at the GitHub repository of the Hybrid Exoskeleton Glove project:

  https://github.com/newdexterity/Hybrid-Exoskeleton-Glove/tree/master/Code/Muscle-Machine-Interface
  

• A Wearable Muscle-Machine Interface Based on Forcemyography and Electromyography

  Lucas Gerez • 10/03/2020 at 18:30 • 0 comments

  In order to control the proposed motorized exoskeleton glove, an Electromyography (EMG) and Forcemyography (FMG) based wearable sleeve system was developed (see Fig.1). The sleeve interface is equipped with 3 bipolar EMG channels and 5 FMG sensors. It is made out of a breathable and stretchable fabric and can be easily worn using a zipper. The FMG sensors are implemented using Force Sensitive Resistors (FSR) and silicone based supporting pads, while the EMG sensors are developed using reusable wet silver electrodes supported by thick silicon blocks to maintain a tight contact with the human skin. The EMG electronics include four stages: i) the differential amplification, ii) band-pass filtering, iii) full-wave rectification and iv) calculation of the envelope of the signal. 

  The FSR sensors are the 402-Round sensors (Interlink Electronics, Camarillo, CA, USA) and have a force sensitivity range of 0.2N-20N which is enough to detect even the slightest muscle movements. The reusable electrodes are manufactured by printing conductive silver ink on poly-ethylene terephthalate (PET) sheets using an inkjet printer. The advantage of using these electrodes over commonly used gel electrodes is that they do not need to be discarded after every use and can be developed in any shape and size to suit the requirements of the application and to improve the efficiency of the system.

  Fig. 2 shows the placement of the FMG and EMG sensors on the human forearm when the sleeve is worn. The sensors E1 and F1 are placed on the extensor digitorum superficialis muscle site to capture the finger extensions, sensor F2 is placed on extensor pollicis brevis muscle to capture the thumb extensions, sensor E2, E3, F3 and F4 were placed on the flexor digitorum superficialis muscle site to capture finger flexion and sensor F5 was placed on flexor digitorum profundus muscle site to capture the flexion of the distal joints when a fist is made.

  The video below demonstrates the operation of the wearable sleeve system:

• Molding Silicones and Urethane Rubbers

  Lucas Gerez • 10/01/2020 at 21:55 • 0 comments

  3D Printing Components

  The rigid parts and molding components of the glove are produced through 3D printing, and are
  printed in PLA or ABS plastics.

  Mold Assembly

  Depending of the module being fabricated, the mold assembly steps can vary from preparation of a
  single mold part, to the assembly of several mold components.
  Before Assembling the components of the mold together spraying the inside of the mold with mold
  with mold release can greatly ease in the effort required during the demolding process.
  The mold release we use is EASE RELEASE MANN 200 FOR SILICONES & RESINS.

  Filling Molds

  Once a mold is assembled, the selected elastomer material is mixed together following the
  manufacturers guidelines and then degassed to remove any air bubbles, which can leave unwanted
  cavities in the molded part. When the mixture is complete it is deposited into the mold cavity,
  where the mold release was also sprayed.

  The chosen elastomers we used to produce the laminar jamming finger pouches, layer jamming material, and telescopic thumb were Dragon Skin 30, and Vytaflex 40. The 2 materials both had a 1 to 1 mix ratio. For more complex parts like the telescopic thumb, which uses a multi-mold process the ‘3D Printing’, ‘Mold Assembly’, ‘filling Mold’, and ‘Mold Disasembly’ steps need to be repeated to mold the different parts together. An Image of the molding steps of the telescopic thumb can be found below. 

  An example of mold mixing can be found in the video below.

  Mold Disassembly

  After the curing process is completed, the molded components can be removed from the molds and
  assembled together with the other 3D printed parts and the glove.

• Results: Body Powered Exoskeleton Glove

  New Dexterity • 07/06/2020 at 00:21 • 0 comments

  The experiment focused on wearing the devices to evaluate the grasping performance for different everyday objects. The goal of these tests was to verify if the subject was capable of executing different grasping tasks and handle the objects as a healthy individual. When the device is locked in a position of grasping, it is possible to grasp and retain the object without difficulties due to the constant force applied by the differential in the body-powered device. In order to evaluate the amount of force exerted by the devices, two different grasping types were used: the pinch grasp and the power grasp. A Biopac MP36 data acquisition unit (Biopac Systems, Inc., Goleta, California) was used with the SS25LA dynamometer to measure the forces exerted in each scenario. The maximum force obtained for the pinch grasp and power grasp configurations were 8.2 N and 11.6 N, respectively. These pinch and power grasp forces are enough to execute most of the activities of daily living.

  The video below demonstrates the operation of the Body Powered Exoskeleton Glove:
  

• Results: Hybrid, Motorized Exoskeleton Glove

  New Dexterity • 07/06/2020 at 00:17 • 0 comments

  The second experiment focused on measuring the maximum forces that the device can apply to grasp objects. In this experiment, a Biopac MP36 data acquisition unit (Biopac Systems, Inc., USA) was used with the SS25LA dynamometer to measure the forces exerted during pinch and power grasps. EMG sensors were connected to the forearm of the subject to monitor the muscle activity and guarantee that the subject was not exerting any kind of involuntary forces while grasping the dynamometer. During the experiment, the forearm was placed on the table surface to keep the hand still, and the system was actuated until the torque limit of the motors was reached (3.8 N.m). Six trials were recorded and the maximum force obtained was 19.5 N for power grasps and 12.4 N for pinch grasps. The required force to grasp objects during ADLs does not exceed 15 N, and the pinch forces required to execute most of the daily life tasks are lower than 10.5 N. Thus, the proposed soft robotic glove can exert enough force to stably grasp everyday life objects.

  The video below demonstrates the operation of the Hybrid, Motorized Exoskeleton Glove:

• Description: Hybrid, Motorized Exoskeleton Glove

  Lucas Gerez • 07/01/2020 at 23:55 • 0 comments

  A Hybrid, Motorized Exoskeleton Glove with Variable Stiffness Joints, Abduction Capabilities, and a Telescopic Extra Thumb

  The proposed device is composed of two main systems: the soft exoskeleton glove and the control unit. The control unit is composed of five Dynamixel XM430-W350-T motors, two mini 12V air pumps, one 12V vacuum pump, three solenoid valves, a microcontroller (Robotis OpenCM9.04), and a small circuit to control the air pumps. All six tendons are connected to the pulleys of the motors and run though polyurethane tubes that are used for tendon routing from the control box to the soft glove. The ring and pinky fingers are connected to the same motor since these fingers have a supplementary role during object grasping. 

  The soft glove system of the proposed device is composed of a thin, high sensibility glove, a tendon-driven system that consists of six artificial tendons, a pneumatic system that consists of four soft actuators, and five laminar jamming structures. Five plastic tendon termination structures are stitched onto the fingertips of the glove. Soft anchor points have been added in the glove structure for rerouting the tendon, offering better sensibility of the grasped objects than the rigid anchor points. The tendon-driven system has a tendon connected to each of the fingertip structures and an extra tendon that is connected to the thumb's interphalangeal joint region so as to allow for the execution of the thumb's opposition motion. 

  A tendon-driven solution for the thumb abduction / opposition was chosen over a soft actuator based solution, in order to avoid the obstruction of the region between the index and the thumb, as many different grasps types require the object to be positioned in-between the thumb and the index metacarpophalangeal joints (in the human hand purlicue area). The tendons used in the exoskeleton glove are made out of a low friction braided fiber of high-performance UHMWPE (Ultra-High Molecular Weight Polyethylene) and can withstand forces up to 500 N. The soft actuators are used for two different purposes, to allow for the execution of the abduction / adduction motion of the fingers and to increase grasp stability by activating a telescopic extra thumb that provides grasp support. Three pneumatic chambers have been developed with a "V" shape, and they have been fixed in the region in between the fingers to facilitate the execution of the abduction motion of the fingers. The soft actuator was designed to provide active assistance on finger abduction and passive on finger adduction, once the human hand is naturally adducted. 

  The soft actuators that have been designed are described in the following subsections. At the back of each digit, laminar jamming structures are attached to control the force required to close the digits, to maintain the fingers steady in a desired configuration, and to perform passive extension of the fingers keeping the hand in its natural, zero effort position. The laminar jamming structures can achieve multiple stiffnesses by applying a pressure gradient into the system and relying on the friction between the layers. A single vacuum pump is used to jam the layers of all fingers, enabling variable joint stiffness. A flex sensor was placed at the index finger region and is used to trigger a desired function of the exoskeleton glove when a set bending angle is achieved. 

  The operation of the device is straightforward. Using the smartphone app, the user selects the mode desired to control the exoskeleton glove. The user can combine the motions (e.g., full grasp with abducted fingers or tripod grasp with the extra thumb inflated). A flex sensor can be selected to trigger the desired motion when a set bending angle is reached. The information is transmitted to a microcontroller through Bluetooth communication. Then, the microcontroller activates the chosen actuators that are connected to the glove.

• Description: Body-Powered Exoskeleton Glove

  Lucas Gerez • 07/01/2020 at 23:53 • 0 comments

  A Body Powered Exoskeleton Glove

  The body-powered exo-glove was designed to enhance the grasping capabilities of the user, providing easiness and intuitiveness of operation, with long autonomy, low maintenance, and low cost. The device consists of four different parts: the differential module, the soft glove, the tendon tensioning and adjustment mechanism, and the harness (see Figure below).

  The differential module is a solution for tendon tensioning and even distribution of the grasping forces for the participating fingers. The particular differential mechanism can also be applied to different underactuated prosthetic and orthotic systems. Differentials based on the whiffletree mechanism are widely used in underactuated robot hands. The differential is divided into three different parts: the ratchet clutch, the linear ratchet, and the spring loaded whiffletree mechanism. The ratchet clutch mechanism consists of a ratchet-pulley block for tendon wrapping, a pawl that blocks the rotation of the ratchet in one direction and an elastic element that acts as a spring and pushes the pawl against the ratchet teeth, constraining its motion in the other direction. This mechanism allows a fine and precise adjustment of the tendon length (with a precision of 0.87 mm). The purpose of using this mechanism is to adjust the length of multiple tendons that are routed through the tendon routing tubes and reach the glove. In order to keep the tendon tensioned for a long time, a linear ratchet was used. This mechanism guarantees that the tendon is locked in one position until the mechanism is used again. This mechanism consists of several "V" shape teeth arranged on a row, a lever, a rail, a base, and two springs. The rail is fixed to the differential module through screws and the base can slide on the rail guaranteeing that the motion of the base always happens on a single axis. When the upper cable is pulled, the lever is pushed by a spring against the teeth until the system reaches the desired position. Then, the lever slides into one of the "V" shape teeth locking the mechanism and keeping the tension constant. When the system is re-engaged, the lever is pulled again to the channel and a spring that connects the base to the differential module walls pulls the base until the lever reaches its lowest position and the tendon returns to its initial tension. When the upper cable is pulled again the cycle is reinitialized. The ability to keep the tendons tensioned for long periods of time is of paramount importance for underactuated and body-powered systems, since in other tendon-driven, motorized solutions (e.g., fully-actuated systems) the dedicated motors can adjust the tensioning of the tendons and hold the load while the grasping and manipulation of the objects take place.

  The body-powered mechanism allows the transmission of forces from the upper body (e.g., the shoulders) to the index, middle, and thumb fingers through the tendon routing system. Simple body movements can increase the tension of the tendon, actuating the soft exo-glove. The differential mechanism is used to evenly distribute the forces to the fingers. In order to operate the device, this cable must be accurately tensioned and for this purpose, a tension adjustment mechanism was designed. The mechanism consists of a base where the parts are connected, a lever, a pulley with rectangular teeth, a cover and a retractable reel. After wearing the mechanism, the user presses a lever and the cable on the reel (separate from the tendon) rotates the pulley in the counterclockwise direction wrapping the actuation tendon around it and tensioning it. The plastic cover guarantees that the cable does not slip out of the pulley channel.

  The proposed harness was chosen for the body-powered device because it is comfortable and helps to keep the shoulders aligned. When the right arm or the shoulders move transmitting forces to the main cable, the differential is pulled and the artificial...

  Read more »

View all 13 project logs