**4. Methodology**

While users walk, the weight they bear on the cane keeps changing. To calculate the maximum support that a given user needs while walking, gait cycles can be analyzed using the adapted cane. Figure 6a shows the typical cane movements for contralateral users while they walk. During a gait cycle, maximum force is applied on the cane when it is fully vertical, as the load vector is orthogonal to the force sensors plane. The cane support cycle corresponds to the elapsed time from a heel strike

of the opposite leg, with respect to the cane, to the next one. As expected, force sensors outputs fluctuate while users change their load on the cane (Figure 6b). As commented, maximum peak values correspond to the cane in a vertical position. A peak during a cane support period represents the upper bound support that a given user needs in each step. A sequence of peaks in time provides continuous information about the user load on the cane. We have used the *findpeaks* MATLAB R2016b (The MathWorks, Inc., Natick, MA, USA) function to detect all peaks during a given test. Users load on the cane during the affected leg step, but then they move the cane. The minimum time between two cane supports depends on the users' step times. More than 99,99% of the elderly population reported a value above 0.5 s [29]. For these reasons, the *findpeaks* function has been set with a *MinPeakDistance* parameter equal to 25 (0.5 s). An additional function parameter, the *MinPeakHeight*, has been defined. It is used to filter out spurious peaks from the input signal. Its value depends on users' load, so we have empirically set it to the average of the input signal exercise.

**Figure 6.** Contralateral user walking: (**a**) User walking in the support cane period; and (**b**) Input signal from force sensor during a 10 s interval in user 8 (range is 0–1023). Maximum peaks (corresponding to cane vertical position) are marked in -.

Our tests were carried out in two senior centers in Cordoba, Spain. As commented, we chose volunteers who require a cane for mobility. We chose to select people with different disabilities that required the cane either for balance or to reduce pain in limbs after surgery or due to degenerative conditions. We also chose people who favor, left, right, or no particular side when walking. Thus, we can check how the module responds to different load bearing profiles. In the end, we selected 10 volunteers: 8 men and 2 women. An additional participant who was initially eligible was excluded from the study, as the module showed that she did not load any weight while she walked, i.e., did not require weight support assistance. Participants were on average 83.7 years old (range 74–94 years). Table 1 shows their ages, genders, average gait speeds, and their physical diseases.

**Table 1.** Condition and characteristics per users.


Volunteers were asked to complete two different popular clinical tests sequentially. All cane data was collected by a mobile phone paired with the cane via BT. First, volunteers performed the 10 m test (Figure 7). Then, they kept walking during 1 more minute approximately, depending on their condition. All tests were approved by the University of Malaga Institutional Ethical Committee and by the senior centers. Also, all volunteers signed an informed consent.

**Figure 7.** Cane user walking during a free walking test.
