*4.1. Methodology*

The pilot study was conducted with a single participant, an active male (K4, 69 years, 1.88 m, 90 kg) with transtibial amputation on the right side (stump length 11 cm, amputated for five years, Pro Flex XC foot (Össur hf, Reykjavik, Iceland)) who is wearing a custom-built measurement socket based on a hydrostatic plaster impression with pin lock mechanism and size 30 silicone RelaxTM Locking liner (Össur hf, Reykjavik, Iceland). Prior to the experiment, the participant provided his written informed consent. The study was conducted with a positive vote by the ethics committee of the Technische Universität Darmstadt (reference number: EK40/2017) and is in accordance with the Declaration of Helsinki in its current version.

The developed measurement system was used to record the relative motion at seven specific socket locations. The sampling frequency was set to 100 Hz per sensor unit. Measurement sites were identified via semi-structured interviews with the participant and the prosthetist. Four of the chosen sites have previously been experienced as problematic and can mostly be assigned to anatomical landmarks (e.g., distal end of residual bones). In addition, three unproblematic socket locations were chosen for comparison. Table 1 summarizes the positions of the individual sensor units.


**Table 1.** Identified measurement sites.

For a seamless sensor unit integration into the socket, the identified sites were already taken into account during the socket manufacturing process: The positive residual limb plaster model with marked locations was digitized with the 3D-scanner ATOS III (GOM GmbH, Braunschweig, Germany), individual sensor base plates corresponding to the plaster model surface at the measurement locations were designed, and 3D-printed in PLA using the Ultimaker 2+ (Ultimaker B.V., Geldermalsen, Netherlands). These sensor base plates were attached to the plaster model, on which the carbon socket was then fabricated.

In addition to the presented measurement system, a custom-built measurement adapter [18] integrated into the prosthetic structure was used to measure the loads at the distal end of the socket with a sampling frequency of 500 Hz. The two systems were synchronized by a triggering signal provided by the adapter. Figure 4 shows the participant wearing the custom-built measurement socket with integrated sensor array as well as the adapter for measuring the occurring loads.

**Figure 4.** Participant equipped with custom-built measurement socket and load adapter.

For this pilot study, the participant completed a 5 m walkway of straight level walking 15 times. Data of the two measurement systems were interpolated to 200 Hz. The measured force in proximodistal (pd) direction was used to automatically identify and isolate completed gait cycles (GC). Of these gait cycles, the first and last steps of each measurement were excluded. In total, 54 gait cycles remained for data analysis.
