2. Exo-Glove II with Dual-Tendon routing

2.1 Implementing Dual-Tendon Routings to Exo-Glove

Fig. R2.1. Derivation of possible dual-tendon routings in the two-linkage system

(a) shows the overall view to derive possible dual-tendon routings for two-finger applications. (b) - (h) show schematics of the dual-tendon routings found for two-finger applications. In these figures, alphabet a, b, c, and d with blue color mean the tendon section at the motor side; numbers 1, 2, 3, and 4 with blue color mean the tendon section at the glove side. (i) shows a description of the components used in the figure.

We can obtain 7 (=2^3-1) DT routings for the Exo-Glove because it is designed to assist two fingers.
Among 7 DT routings (TR1 - TR7), the most appropriate tendon routing has been obtained below.

2.2 Finding the best "dual-tendon routing" for Exo-Glove among TR1 ~ TR7

Fig. R2.2. Tension distribution of possible DT routings

Table. R.2.1 Performance of the dual-tendon routing

By measuring the tension at the end-effector sections, we can obtain the tension distribution of TR1 - TR7 (Fig. R.2.2.)
Using the five performance factors (see our paper for more details), we conclude that TR3 is the most appropriate tendon routing for the Exo-Glove.

2.3 Result - Improved performance of the Exo-Glove II (by using TR3)

Fig. R2.3.1 Pictures of grasping objects with the Exo-Glove II

Fig. R2.3.2 Two possible methods to measure the tension.

2.3.1. The use of TR3 improves the under-actuation performance"
(Measured by the difference between two fingertip forces)

Robot with the previous routing (TR2): 3.85N
Robot with proposed routing (TR3): 1.97 N

2.3.2 The use of TR3 reduces hysteresis at the flexor.
(Measured by the restored joint angle after actuation)

Robot with the previous routing (TR2): returned to 63.27% of ROM
Robot with proposed routing (TR3): returned to 27.91% of ROM

2.3.3 The use of TR3 enables the usage of a passive tendon.
(Measured by the required tension to extend the fingers after actuation)

Robot with the previous routing (TR2): requires 5.56 N to extend the fingers
Robot with proposed routing (TR3): requires 3.85 N to extend the finger.

2.3.4 The use of TR3 enables attachment of the compact tension sensor at the wearing part

A robot with the previous routing (TR2): should use the bulky tension sensing method shown in Fig.R2.3.2 (a). Three bearings and corresponding fixing-parts make the system bulkier.
Robot with proposed routing (TR3): can use compact tension sensing method shown in Fig.R2.3.2(b).

2.4 Details of the Experimental Setup

Fig. R2.4.1 Experimental Setup for Measuring Tension Distribution. (a) The tendon tension at end-effector side is measured using load cells. (b) Bowden cables are positioned between the motor side and the finger side to replicate the conditions of a real robot. In the experiment, the cables were arranged to bend 360 degrees. (c) and (d) illustrate tendon routing examples used in experiments TR2 and TR6, respectively. (e) The motor-side wire path holder allows for seven distinct tendon routings without altering the overall experimental setup. This component includes four pulleys that can function as movable pulleys for Dual-Tendon Routings. (f) A photograph of the experimental setup is shown.

Fig. R2.4.2 Experimental Setup for Measuring Robot Performance. (a) Experimental setup for measuring the fingertip force of the index and middle finger, with results are explained in section III.B of the main text. (b) Experimental setup for evaluating the position hysteresis. Ten markers on the glove track joint angle, while two markers on the actuator measure the actuator displacement. Results on positional hysteresis are presented in section III.C and III.E.

Components Used in the Experimental Setup

Load Cell: 333FB Cell, Ktoyo Co., Ltd., Korea
Pulleys for Tension Measurement: 18 mm diameter pulleys
Bowden Cable: Custom-made compression spring (outer/inner diameter: 3/2.5 mm) paired with a Teflon tube (outer/inner diameter: 2/1 mm)
Motor and Motor Driver: DC-Micromotor 2232 SR with a gear ratio of 69:1, controlled by a Motion Controller (MC 5004P) from Faulhaber

Page updated

Google Sites

Report abuse