3.4.1. Exoskeleton Testing without Wearer's Hand

The mechanical behavior of the exoskeleton was analyzed using Kinovea video processing software. The trajectories of key points were tracked via a high frame-rate camera for determining the workspace of the exoskeleton finger. The procedure used to capture and process the video via Kinovea software was described in a previous paper where the software was used to study and determine the anthropometric parameters of the human hand [53]. In Figure 8a the key points and their notations are illustrated. After studying the mechanical behavior of the system, it was observed that the joints rarely moved simultaneously in relation to one another; this phenomenon is simply explained by the friction differences from one joint to another and the friction variable of the cable on the contact guiding surface. As a result of this phenomenon, the joints of the exoskeleton will move sequentially one joint at a time. This phenomenon does not constitute a disadvantage since it offers a good indication for the positions where the system encounters greater frictions and it can be traced back to optimize the mechanism.

*Symmetry* **2020**, *12*, x FOR PEER REVIEW 9 of 17

**Figure 8.** Joint phase steps of one flexion/extension cycle. (**a**) Key points and their notations. (**b**) Initial state of the exoskeleton. (**c**) The first phase of actuation—DIP joint flexion movement (**d**) The second phase of actuation—PIP joint flexion movement. (**e**) The third phase of actuation—MCP joint flexion movement. (**f**) The first phase of actuation—PIP joint extension movement. (**g**) The second phase of actuation—PIP joint extension movement. (**h**) The third phase of actuation—MCP joint extension movement. In Figure 8 the phases of a flexion/extension cycle are detailed. In Figure 8c, it is observed that **Figure 8.** Joint phase steps of one flexion/extension cycle. (**a**) Key points and their notations. (**b**) Initial state of the exoskeleton. (**c**) The first phase of actuation—DIP joint flexion movement (**d**) The second phase of actuation—PIP joint flexion movement. (**e**) The third phase of actuation—MCP joint flexion movement. (**f**) The first phase of actuation—PIP joint extension movement. (**g**) The second phase of actuation—PIP joint extension movement. (**h**) The third phase of actuation—MCP joint extension movement.

on the same logic, the extension's phase steps will be dependent on each joint's friction. As seen in Figure 8f, the first joint in the sequence to move is the PIP joint, followed by the DIP, as seen in Figure 8g. The final movement is again the MCP remote center of rotation, as seen in Figure 8h. An essential aspect to point out is that although the order of the joint movements for this cycle of flexion/extension is as expected, the system is still an underactuated mechanism. Given the right conditions, the order of the joint movements will not always be the same. Small variations of friction from the joints or cable can result in a different order in the joint movements, producing an asymmetric trajectory cycle during flexion/extension exercises. The next step in the analysis is generating the trajectory of the key points and the exoskeleton's workspace. Based on the captured motion of the key points over several

flexion/extension exercises, the workspace seen in Figure 9 is generated.

in the first phase of the finger actuation, the DIP joint is the first to move. The movement of the PIP

In Figure 8 the phases of a flexion/extension cycle are detailed. In Figure 8c, it is observed that in the first phase of the finger actuation, the DIP joint is the first to move. The movement of the PIP joint, as seen in Figure 8d, characterizes the second phase. The third phase is the rotation of the MCP remote point, as seen in Figure 8e. The behavior observed is as expected, since the DIP and PIP joints each have two ball bearings, while the MCP remote center of rotation utilizes a more complex mechanism with multiple joints that will inherently encounter more significant friction forces. Based on the same logic, the extension's phase steps will be dependent on each joint's friction. As seen in Figure 8f, the first joint in the sequence to move is the PIP joint, followed by the DIP, as seen in Figure 8g. The final movement is again the MCP remote center of rotation, as seen in Figure 8h. An essential aspect to point out is that although the order of the joint movements for this cycle of flexion/extension is as expected, the system is still an underactuated mechanism. Given the right conditions, the order of the joint movements will not always be the same. Small variations of friction from the joints or cable can result in a different order in the joint movements, producing an asymmetric trajectory cycle during flexion/extension exercises. The next step in the analysis is generating the trajectory of the key points and the exoskeleton's workspace. Based on the captured motion of the key points over several flexion/extension exercises, the workspace seen in Figure 9 is generated.

**Figure 9.** Exoskeleton key point asymmetric trajectory cycle and workspace analysis without operator's hand.

A more detailed analysis of the mechanical behavior is illustrated in Figure 10a, where the processed video data is used to generate a graphical representation of the laws of motion for each key point on the X and *Y*-axis.

**Figure 10.** (**a**) Graphical representation of the law of motion of each key point when the wearer's hand is not mounted on the exoskeleton. (**b**) Influence of the joints' actuation order on the law of motion and torque.

The chosen example for generating the graph in Figure 10a, confirms that not all flexion/extension cycles have the same order of actuation of the joints when the human hand is not interfaced in the exoskeleton. It is observed that the first four cycles follow the sequence of steps described in Figure 8, while the next three cycles show a noticeable difference, which is the result of a different order of the joints' actuation. The difference is at the flexion part of the cycle, where the first four cycles follow the order DIP, PIP for flexion, and then MCP, while the remaining cycles follow the order DIP, MPC, and then PIP. This change in the joints' order of actuation is also observed in torque, as seen in Figure 10b, where the torque is significantly smaller after the change in the joints' actuation order.
