3.4.2. Exoskeleton Testing with Wearer's Hand

Similar to previous tests, a trajectory was determined for the key measurement points. In this test, the human hand is interfaced with the exoskeleton. The generated path can be observed in Figure 11, where seven flexion/extension cycles were used to generate the data. An important observation is that, compared with the previous experiments, the mechanical behavior, such as the order of joint actuation, has changed. The first thing that was noticed is that the trajectory of the key points has changed in a way that the flexion/extension exercises tend to produce a more symmetrical pattern with the wearer's hand in the exoskeleton. While experimenting without the wearer's hand in the exoskeleton, the joint actuation was generally done one joint at a time. With the wearer's hand in the exoskeleton, the joints are actuated simultaneously, most of the time. This behavior is due to the reaction forces and friction forces introduced in the system by the biological finger, and its interactions with the exoskeleton's orthotic shell. In Figure 12a detailed graphical representation of the law of motion of the key points with the operator's hand in the exoskeleton is presented.

**Figure 11.** Exoskeleton key point trajectory analysis with operator's hand.

**Figure 12.** (**a**) Graphical representation of each key point's law of motion, while the operator's hand is mounted. (**b**) Torque variations as resulted from operator finger alignment in the exoskeleton.

A comparison of the cable displacement and torque is made for the seven cycles of flexion/extension, as seen in Figure 12b. Another interesting phenomenon that appeared in this experiment is that the torque values measured decreased after a random number of cycles, as seen in Figure 12b; this phenomenon was investigated further to determine the cause. After extensive experimenting and comparing the data, it was observed that this phenomenon happens in almost every experiment iteration with the wearer's hand in the exoskeleton.

As seen in Figure 13, the torque and also cable displacement start to display a more regular pattern after 200 measurement samples. Extensive testing concluded that the decrease in torque occurs due to the finger self-alignment in the orthotic shell after several flexion/extension cycles. Although the mechanism functions, the biological articulation of the finger and the mechanical joints of the exoskeleton do not match perfectly at the start of exercise, due to coaxial offsets. This phenomenon produces an asymmetric behavior at the start of the exercises and after the biological and mechanical axis auto-align, the behavior tends to become symmetrical. This logical explanation corresponds to the irregular behavior of the system at the beginning and during the first flexion/extension cycles.

**Figure 13.** Torque and cable displacement stabilization due to the biological finger self-alignment in the orthotic shell of the exoskeleton.
