Body flexibility of MR-LF-S

As noted in the text, the closeness of the MR-DF and MR-LF curves indicates that the method of localizing flexibility was successful at reducing the bending region length while preserving body flexibility. The data for the MR-LF-S (shown in Figure \ref{234649}A) has a higher amplitude, indicating higher flexibility. Further, the flat section at the top of the MR-LF-S curve appears to be a result of the foot contacting the compartment. Indeed, images show the MR-LF-S foot contacting the body at the maximum and minimum flexion (indicated by red arrows in Figure \ref{234649}B), which may cause undesirable effects in locomotion. Because the geometry of the flexure was the same between MR-LF and MR-LF-S models, the additional bending in the MR-LF-S model could be due to bending in the soft compartment and an expected difference in wall effects compared to MR-LF.
 
A comparison of locomotion results between MR-LF and MR-LF-S with the same mass show that MR-LF-S exhibited a slightly higher initial speed (4%) at ya = 11 cm, slower initial speed (47%) at ya = 15 cm (Figure \ref{618954}A), and a longer period of inhibited gait for locomotion of one trial at ya = 10 cm (indicated by right-pointing arrow in Figure \ref{618954}B), which was not included in da*. The data in Figure \ref{618954}B also show that the mean of data within da* was close between robots (MR-LF: 14.22 cm, MR-LF-S: 14.19 cm), indicating that the robots stopped locomotion at a similar location, likely as a result of their equal mass. Additionally, the MR-LF-S da* had a larger standard deviation (MR-LF: 0.23 cm, MR-LF-S: 0.75 cm), indicating MR-LF-S had more variability in locomotion distance after 40 steps. From these results, we see that a robot with a soft compartment (i.e., MR-LF-S) exhibits comparable locomotion to a robot with a rigid compartment (i.e., MR-LF), where differences between the designs can be attributed to the additional flexibility (increase of 61% in maximum and 48% in minimum foot flexion) exhibited by the soft body configuration. These observations can be applied in future applications when, for example, fabricating entirely-soft robots by magnetic 3D printing.