The protocol for this phase of the study began with the completion of a preliminary questionnaire and involved individual pressure map measurements. Upon arrival at the facility, each participant reviewed and provided written informed consent for their voluntary participation. This process was conducted following Mexico’s General Health Law, ensuring compliance with the ethical principles outlined in the Declaration of Helsinki. Considering the minimal risk associated with the study, the protocol was subjected to an internal review conducted by a panel of institutional researchers.

Each participant completed a questionnaire specifically designed to gather relevant personal information, including daily routines, date of birth, and occupation. After confirming that participants met the established inclusion and exclusion criteria, anthropometric measurements were taken. Initially, each participant was weighed, and their height was measured to calculate their Body Mass Index (BMI). Those who did not meet the BMI-based inclusion criteria were excluded from the study. For participants who met the criteria, additional anthropometric measurements were recorded. The participants were instructed on the procedure for taking these measurements, which were conducted while standing in a bipedal stance. Measurements of the torso, hip, and thigh were performed according to the standardized protocols of the International Organization for Standardization (International Organization for Standardization, 2017). Similarly, sitting dimensions were recorded, including torso height, thigh length, and the distance between the knee and the foot support.

After completing the collection of anthropometric data, the pressure measurement station was prepared. The station consisted of a chair with a backrest and a flat seating surface. A 15 mm-thick Medium-Density Fiberboard (MDF) panel was placed on the chair seat to ensure a uniform contact surface and standardize the measurement base. Finally, the CONFORMat® sensor matrix (Tekscan® Inc., Norwood, Massachusetts, USA), measuring 45.44 x 45.44 cm, was positioned on top of the MDF panel. This entire system was utilized for conducting pressure measurement tests, as shown in Figure 2.

After installing the sensor matrix, the subject’s anthropometric weight data was entered into the software. The measurement equipment was then calibrated according to the user manual and prepared for data collection while the subject was seated in the chair equipped with the sensor matrix, as shown in Figure 3 and Figure 4. This procedure was essential, as the sensor system was customized for each new participant; every measurement was tailored to the individual’s height and weight. The equipment was calibrated, and participants were instructed to assume three distinct postures: an upright torso, a torso without lumbar support, and a forward-leaning torso, as shown in Figure 4.

Each trial for data acquisition had a duration of one minute. To guarantee the reliability and representativeness of the data, three trials were undertaken for each posture, culminating in a total measurement duration of nine minutes per participant. The information obtained from the measurements in each posture was represented as a two-dimensional pressure map, as is shown in Figure 5.

Following the collection of pressure maps from participants, a reference database was established to facilitate statistical analysis. The outcome variables included peak pressures at the ischial tuberosities measured in kPa; average pressure across the contact surface also measured in kPa; mean contact area measured in cm²; and average pressure maps (Summary of all variables is presented in Table 1).

Hypothesis testing was conducted to assess the influence of posture on pressure distribution and to ascertain whether specific usage recommendations for the cushion were warranted. An initial analysis of variance (ANOVA) revealed a lack of normality in the data for the “Upright Torso” posture. Consequently, the analysis was supplemented by employing the non-parametric Kruskal-Wallis test. An examination of the data distributions, in conjunction with the results from the Kruskal-Wallis test, confirmed that there were no significant differences (p>0.05) in pressure metrics among the three postures. Statistical analyses were conducted using MATLAB® software (version R2015a, MathWorks®, Natick, MA, USA).

The database was carefully organized to consolidate all relevant information. Initial statistical estimations were performed, which included calculating the mean, variance, and confidence intervals. Additionally, both parametric and non-parametric tests were applied to compare the means. To create the average pressure maps, an XLSX file from the CONFORMat® software was exported. This file included the pressure values recorded by each sensor, organized into a 32 x 32 cell matrix that reflected the layout of the sensor configuration. To obtain the average pressure maps, the mean of the matrices from all forty participants was calculated, and the resulting matrix was graphically represented using MATLAB® software. Based on the insights gained from these maps, various geometric and material configurations for the ergonomic cushion design were developed, guided by the average pressure patterns associated with each evaluated posture.

The Taguchi method was applied using an L9 orthogonal array as shown in Table 2, designed to examine the influence of three design factors, each at three levels (Gutiérrez & Salazar, 2012). This methodological approach enabled the assessment of how geometry, thickness, and material composition affect pressure distribution, thereby facilitating the identification of optimal design combinations.

Table 3 shows the nine unique combinations of factor levels used in the experimental runs, as specified by the Taguchi orthogonal array design.

The Signal-to-Noise (S/N) ratio was calculated for each factor combination using the “Less-is-better” criterion. This choice was based on the goal to reduce the pressure on the contact surface and, as a result, to find factor levels that lower variability in pressure distribution. By analyzing the S/N ratios, the main effects of the design factors were identified, leading to the best combination of factor levels for cushion design. The cushion based on this optimal setup was then compared with other designs to assess its relative performance.

As previously mentioned, since no significant differences were found among the three postures studied, the decision was made to use the pressure map from a single posture as the basis for designing the cushion geometry models. This selected posture was “Upright Torso,” as it exhibited the highest peak pressure among the three. Based on this, three cushion designs were modeled using Autodesk® Fusion 360 software (version 2.0, Autodesk® Inc., San Rafael, CA, USA), as shown in Figure 6.

Finite element analysis is a valuable tool used in engineering to obtain preliminary insights into the behavior of materials or structures. Finite element models are developed to simulate real conditions and predict the performance of each cushion model (Alawneh et al., 2022). The cushion models were discretized into square areas, to which loads corresponding to the average pressure map values for the “Upright Torso” posture were applied, as shown in Figure 7. Boundary conditions were defined to replicate the cushion supporting body weight in the “Upright Torso” posture on a flat surface, as illustrated in Figure 7a. The translational and rotational movements of the cushions were constrained along the global x, y, and z axes on the surface simulating ground contact, as is shown in Figure 7b.

Consistent with the results from the experimental design, physical prototypes of the three identified geometries were fabricated, as illustrated in Figure 8. These prototypes underwent a detailed evaluation by the same group of participants who took part in the initial study phase. This evaluation included both direct physical interaction tests and the completion of a structured questionnaire designed to capture both subjective and objective perceptions. The questionnaire contained open-ended questions aimed at exploring the comfort experience, as well as a ten-point Visual Analog Scale (VAS) that allowed participants to provide quantitative assessments of each prototype, reflecting their personal preferences. The performance of the prototypes was validated by replicating the measurement methodology used in the initial study phase, ensuring that the data were comparable.

A summary of the anthropometric measurements for the 40 participants is presented in Table 4.

As shown in Figure 9, an analysis of the mean contact area for the three postures studied showed that the data followed a normal distribution, as confirmed by the Shapiro-Wilk test. Additionally, the homogeneity of variances was verified using Bartlett’s test. The ANOVA revealed no statistically significant differences in the mean contact areas among the different postures, indicating that the posture adopted did not have a significant effect on the contact area with the seat.

An analysis of mean pressure showed that the highest average pressure was recorded for the “Upright torso without lumbar support” posture, at 12.93 ± 4.31 kPa. The “Upright torso” posture had slightly lower readings, approximately 12.67 ± 4.17 kPa. This indicates that, despite variations in body inclination and arm support, mean pressure remained relatively consistent across these positions, as is illustrated in Figure 10.

A non-parametric Kruskal-Wallis test was applied for a more thorough statistical analysis of mean pressure, as the ANOVA residuals did not meet the normality criterion. The resulting p-value was 0.2697, which exceeds the 0.05 significance level. This led to the conclusion that there were no statistically significant differences in the median pressures recorded for the three evaluated postures.

The data presented in Figure 11 shows that the “Upright torso” posture resulted in the largest mean contact area, measuring 739.61 ± 165.75 cm². In contrast, the “Forward-leaning torso” posture showed the smallest mean contact area at 650.60 ± 157.69 cm². Despite the variations in body posture, the mean contact areas were relatively similar, with only a slight decrease noted when the torso was leaned forward.

The highest concentration of pressure is typically observed over the ischial tuberosities, gradually decreasing towards the edges and front of the seat. While the pressure distribution is not uniform, it shows a general bilateral symmetry, indicating that most participants apply similar pressure on both sides of their bodies when sitting upright. This posture also recorded the highest peak pressures among the three studied positions, as is represented in Figure 12, with values of 40.70 ± 10.06 kPa on the left side and 42.92 ± 18.78 kPa on the right. Peak pressures highlight the areas where the majority of body weight is concentrated. There is a distinct lack of uniformity in pressure distribution, with significant concentrations occurring in the central-posterior region of the seat, where these peak pressures are found.

The maximum pressure at the ischial tuberosities, as shown in Figure 13, was identified as the response variable based on simulations with specific loads.

The results from each simulation, which correspond to the factor combinations defined by the Taguchi Design of Experiments (DOE), were compiled into a dataset shown in Table 5. The following analysis focused on calculating the S/N ratios using the ‘less-is-better’ criterion to identify the optimal factor combination.

The ANOVA results presented in both Table 6, which shows the mean responses, and Table 7, which details the S/N ratios, indicate that both material and geometry are significant factors, as their p-values are below 0.05. This finding demonstrates that the type of material and the geometry significantly influence both the mean response and its variability. In contrast, thickness did not show any statistical significance in this analysis.

In terms of design stability against variations, Figure 14a demonstrates that using a square geometry along with memory foam maximizes the S/N ratio. This combination enhances the robustness of the optimal performance configuration by minimizing variability. Figure 14b highlights that the ideal setup for average performance consists of a square geometry (level 3), memory foam (level 1), and a thickness of 12 cm (level 3). While thickness did not show statistical significance, the graph indicates its role in maximizing overall performance.

The comparative analysis with the physical cushions shows, as illustrated in Figure 15b, c, and d, that there is a reduction in peak pressures in the ischial tuberosity regions. Additionally, there is a significant increase in contact area while maintaining a similar distribution on both sides.

The optimal cushion identified through the Taguchi method increased the contact area by 86% to 103% compared to the initial baseline measurement, as detailed in Table 8.

The prototypes decreased peak pressures in the areas of highest initial concentration by over 40% compared to the maximum pressure initially observed on the tissues, as shown in Table 9.

The results from the experimental validation and numerical simulation tests showed strong agreement, with differences of less than 3% for the peak pressures of the left and right ischial regions as presented in Table 10.

The satisfaction survey using a Visual Analog Scale (VAS) conducted during prototype validation showed that the ‘Square’ design received the highest ratings for stability and comfort, as shown in Table 11, effectively meeting user expectations.