Despite the interest in under-actuated tendon-driven (UATD) wearable robots, their practical issues have not yet been addressed. 

• Friction at the tendon is significant as it accumulates when the tendon passes many joints and fingers. Friction causes detrimental effects on tension distribution, hysteresis, reliability, and efficiency.

• Tendon Elongation is big because the tendon length is too long. Therefore, the elongation gets bigger in wearable robots because the actuator is located far from the wearing part. Tendon Elongation makes it difficult to control the position of the end-effector. It also increases the actuator size because the actuator should pull more tendons.

• These issues can be easily solved in conventional robots by using bearings and thick tendons. However, using these components is NOT preferred in "soft wearable robots" 

• Alternatively, we can solve the issues by finding a better tendon routing.
  *Tendon routing means "how the tendon passes the joints and fingers" 

However, there exist "many" ways to route the tendon 

in the same configuration (with the same joints and fingers).

This paper has two contributions (Please click the title of each contribution for the detailed content): 

• Dual-Tendon Routing (DTR) for soft hand wearable robots [Figure R.O1(b)]
  The paper introduces an under-actuated tendon routing, called Dual-Tendon (DT) routing, to solve the Friction and Tendon Elongation issues. The paper also presents a method to find 2^(n+1)-1 DT routings for robots that assist n fingers.  Researchers can choose the most appropriate tendon routing among the obtained DTR sets.
  
  

• Exo-Glove II design with DTR [Figure R.O.1(c)]
  The Framework for DTR is applied to Exo-Glove, a hand-wearable robot developed for spinal cord-injured people. Among 7 DTRs, the most appropriate routing is chosen using five performance factors (adaptability, torsional balance, reliability, efficiency, and transmission ratio). The use of this framework 

1) improves under-actuation performance.

2) reduces hysteresis at the flexor.

3) replaces the active extension tendon with the passive tendon (since the friction at the flexion tendon is dramatically reduced).

4) enables the tension sensor to be attached to the wearing part in a compact size.