Mechanical optimization of a rehabilitation harness via an integrated Abaqus-ISIGHT framework: A multidisciplinary approach for enhanced biomechanical performance

Abstract

Medical-device-related pressure injuries resulting from mismatched human–harness interfaces during body weight support (BWS) training represent a persistent challenge in clinical rehabilitation. This study investigates a multidisciplinary cosimulation framework integrating finite-element analysis (FEA) and optimization algorithms to enhance the biomechanical compatibility of harness systems. An automated multisoftware optimization workflow was developed to minimize stress and strain induced by the harness, with the objective function targeting optimal pressure distribution across the body–harness interface. Subsequently, inverse parameter identification was performed to derive the optimal material properties of the harness. Simulation results revealed that the optimized harness configuration significantly reduced peak von Mises stress by 15.47% ± 1.2% and maximum shear strain by 19.58% ± 0.9% compared with conventional designs. Furthermore, a region-specific material optimization strategy was implemented: stiffer, warp-dominant materials were employed in the lumbar region to enhance load-bearing performance, while softer materials with lower warp and weft stiffness were applied to the shoulders and legs to redistribute mechanical loads and reduce localized tissue deformation. These findings provide a biomechanically informed design strategy that effectively mitigates the risk of pressure injuries during BWS training, contributing to safer and more comfortable rehabilitation harness systems.

Keywords

Human body motion human sensory and comfort issues materials measurement performance product design systems product and systems engineering

The body weight support (BWS) system is a widely utilized lower-limb rehabilitation robot employed in gait training for neurologically impaired patients, including those with stroke, spinal cord injury, and Parkinson’s disease.¹ During BWS training, the patient walks on a treadmill while wearing a textile-based harness attached to an overhead suspension system to relieve a portion of body weight,² and an electric lifting motor on the support frame that uses a slip gear and gearbox is responsible for lifting. The primary load-bearing regions of the rehabilitation harness precisely correspond to its anatomical coverage areas, specifically the shoulders, waist, abdomen, and leg, which collectively constitute the human–machine flexible interface that represents a key focus of this investigation.

Compared with the conventional rigid countermeasures, an intrinsically flexible and extensible harness is tightened around the patient’s body segments during the BWS training, forming a soft human–harness interface.³ The relative motion between biological tissues and interface materials generates pressure and shear forces, straining skin and soft tissues. Insufficient force risks displacement, while excessive and prolonged force induces discomfort, reduces compliance, and impedes rehabilitation robotics adoption. Severe cases may occlude capillaries, causing ischemia or pressure injuries.⁴ According to a recent survey, adverse events associated with robotic-assisted therapy can be as high as 17%, with soft tissue pressure injury ranking highest. Borggraefe et al. reported that all soft tissue pressure injuries were caused by straps or cuffs in rehabilitation equipment.⁵ Pressure injuries associated with such medical devices represent a significant challenge in healthcare settings.⁶

The interface between patient and BWS system is one of the key factors influencing how force is distributed on the interface and transmitted from the robot to the body.⁷ As the sole human–robot contact component, the rehabilitation harness sustains up to 80% of the patient’s body weight. Extensive research has demonstrated that the anisotropic properties and dynamic tensile mechanics of fabrics are closely associated with wear pressure comfort.^8–10 The ergonomic design of the harness, such as fabric structure and material properties, is crucial for the patient’s well-being and rehabilitation performance. Integrating inherent safety principles or safety-by-design strategies can mitigate these risks early in the design process. However, existing knowledge regarding the optimal structural and material characteristics of the human–harness interface remains limited.

Advancements in textile engineering now integrate 3D virtual fitting systems with comprehensive simulation platforms, enabling accurate prediction of fabric mechanics and real-time visualization of human–harness interaction dynamics.¹¹ Tian et al. demonstrated CLO3D feasibility for wear comfort and interface pressure simulation,¹¹ whereas Hu and Song confirmed its scientific reliability.¹² Though Youn et al. noted CLO3D tends to overestimate localized pressures, the data remain valuable for wearability assessment.¹³

While conventional approaches employing isolated textile modeling or finite-element analysis (FEA) demonstrate fundamental efficacy, their singular implementation fails to address the multidimensional challenges of biomechanical optimization. iSIGHT algorithmic framework is a powerful computational platform for optimizing complex engineering and scientific processes.¹⁴ It integrates multidisciplinary design optimization, process automation, and data-driven decision-making, enabling seamless workflow management and robust design analysis. This computational synthesis facilitates precision optimization of key human factors, such as thermal and moisture comfort, as well as interface pressure management, thereby establishing a novel paradigm for automated textile engineering solutions. Therefore, an innovative digital twin methodology is required: one that integrates textile modeling, FEA, and the iSIGHT algorithmic framework for automated parameter optimization.

To assess human–harness interface compatibility and pressure injury risk, interface pressure is a critical and easily measurable parameter that serves as an efficient indicator of discomfort.¹⁵ Excessive interface pressure is a well-documented risk factor for localized discomfort, restricted blood perfusion, and pressure-related injury. Literature has reported region-specific tolerances for interface pressure in seated and loaded postures. Ciaccia and Sznelwar have suggested specific pressure thresholds, such as 4.25 and 7.8 kPa, beyond which capillaries could become obstructed and tissue oxygenation could be compromised.¹⁶ Mitsuno and Kai found that pressure thresholds could vary depending on the body region (5.8 kPa under the ischia; 2.9 kPa under the thighs),¹⁷ consistent with Kermavnar et al. who showed that discomfort was triggered at mean inflation pressures of 4 kPa at the thighs, 2.46 kPa at the waist, and 1.67 kPa at the shoulders.¹⁸

However, interface pressure cannot reveal mechanical sub-dermal tissue conditions. The predominant theory of pressure injury development involves compression-induced ischemia and excessive shear strain. Studies suggest that the combined effects of cyclic and sustained stress, along with shear strain exceeding a critical threshold, significantly contribute to soft tissue damage.¹⁹ It is therefore essential to assess both the interface pressure as well as internal stress and strain of soft tissue layers.²⁰ Regarding the physiological safety thresholds for the soft tissue stress/strain, an animal study suggested that principal shear strain exceeding 50% for approximately 10 minutes pose an immediate risk of mechanical pressure injury.²¹

This study aims to develop an automated computational framework for optimizing human–harness interface parameters by integrating 3D modeling, FEA, and iSIGHT optimization tools. The framework is designed to systematically reduce interface pressure and mitigate soft tissue stress and strain, addressing biomechanical risks associated with prolonged human–harness interaction. Specifically, this research focuses on two key aspects: (1) establishing a robust numerical framework that enables simultaneous multiobjective optimization of interface pressure and subcutaneous tissue stresses through coupled workflows; and (2) quantitatively analyzing the relationship between critical harness design parameters and their biomechanical performance. The proposed methodology offers a generalized paradigm for optimizing wearable robotic systems through computational mechanics and data-driven parameter exploration.

Description of methodology

The overall research framework diagram is shown in Figure 1. This study systematically investigates two critical parameters influencing the harness system, structural configurations and material properties, aiming to optimize human–harness interface pressure distribution and enhance biomechanical compatibility.

Figure 1.

The framework.

Harness digital design and structural optimization

Digital design

The CLO3D garment modeling platform (CLO Virtual Fashion, Seoul, South Korea) was utilized to conduct physics-based simulations of textile mechanics and real-time dynamic fitting visualization.^12,22 By integrating a digital twin framework, the system facilitates predictive biomechanical modeling of human–harness interactions. The pressure analysis module quantifies interface pressure magnitude and distribution under both static and dynamic loading conditions, enabling precise evaluation of harness fit and mechanical behavior.

Although the human–harness interface plays a crucial role in gait biomechanics during BWS training, it has been underexplored in prior research.²³ The harness, functioning as a human–robot interface, consists of three primary components: shoulder straps, waist straps, and leg straps. As illustrated in Figure 2, four harness designs with distinct structural characteristics were assessed, representing the most used and commercially available harness types.

Figure 2.

Four kinds of harness samples.

On the one hand, these configurations reflect variations commonly observed in existing rehabilitation harnesses, enabling us to evaluate and compare their mechanical performance systematically. On the other hand, each design provides a distinct pattern of load distribution and contact area coverage on the body, influencing how forces are transmitted through the harness.

Based on structural characteristics, four kinds of harness samples were named H-style (H-shape shoulder strap, full-coverage waist strap and parallel-shape leg strap), X-style (X-shape shoulder strap, partial-coverage waist strap, and cross-shape leg strap), Y-style (Y-shape shoulder strap, partial-coverage waist strap, and parallel-shape leg strap), and V-style (H-shape shoulder strap, full-coverage waist strap, and cross-shape leg strap).

Pressure simulation

The CLO3D pressure simulation module was employed to collect biomechanical data from eight anatomical landmarks, covering bilateral shoulders, lumbar regions, and femoral segments. These points were chosen to represent the primary load-bearing and high-contact regions between the harness and the human body.

For the upper body, landmarks were selected along the shoulders and waist where strap loads are concentrated. For the lower body, measurement points were positioned at the thighs, which are the major support and constraint zones across harness.

The measurement points and their corresponding pressure value identifiers included shoulder left (S_L), shoulder right (S_R), waist front (W_F), waist back (W_B), waist left (W_L), waist right (W_R), leg left (L_L), and leg right (L_R).

To standardize shoulder and leg measurement locations (indicated by red markers in Figure 2), shoulder measurements were positioned along the midline of the trapezius region, approximately aligned with the acromioclavicular joint bilaterally. Leg pressure points were defined approximately 3–5 cm below the perineum, targeting the upper medial thigh where the leg straps made direct and stable contact across all harness configurations.

To ensure both safety and comfort, two optimization objectives were established: preventing interface pressures from exceeding critical physiological thresholds (safety constraint), and ensuring uniform interface pressure distribution within physiologically optimal ranges (comfort requirement). Three quantitative pressure metrics were analyzed: peak pressure magnitude (evaluating localized excessive force risks), mean pressure values (assessing overall pressure levels), and pressure distribution variability (standard deviation, quantifying uniformity).

Specifically, peak pressure reflects localized mechanical loading and has been associated with tissue injury risk in pressure ulcer and rehabilitation studies.²⁴ Mean pressure quantifies the average load across the contact surface, serving as a general indicator of overall interface pressure.^25,26 Pressure distribution variability, expressed as the standard deviation of local pressures, reveals the degree of uniformity in load transmission and is used to assess pressure redistribution performance and perceived comfort.^27,28

It should be emphasized that during repeated simulations with the identical harness, interface pressure discrepancies may occur at the same anatomical locations due to harness slippage. Therefore, to accurately represent the overall loading characteristics, the reported pressure data were derived from the averaged values of 10 repeated donning simulations.

Figure 3 presents a comparative heatmap manually generated based on statistical averages of pressure data from multiple wearing simulations in CLO3D. This visualization effectively demonstrates the pressure distribution patterns across the body surface, identifies localized high-pressure concentration zones, and enables systematic inter-harness comparisons within specific anatomical regions.

Figure 3.

Pressure heat maps of the four groups of weight-support harness.

Weight analysis and comprehensive evaluation

To ensure an objective and robust evaluation of harness pressure performance across multiple anatomical regions, this study adopted the Criteria Importance Through Intercriteria Correlation (CRITIC) method to assign weights to 13 evaluation indicators (as listed in Table 1). To comprehensively evaluate the mechanical performance of different harness configurations, in line with the pressure measurement section presented earlier, three key pressure metrics were adopted: peak pressure, mean pressure, and pressure distribution variability. These indicators have been widely recognized in prior biomechanical and ergonomic research for their relevance to comfort, safety, and soft tissue protection.

Table 1.

Description of the 13 evaluation indicators

No.	Indicator	No.	indicator	No.	indicator
1	Pressure test value 1	6	Pressure test value 1	11	Peak pressure value
2	Pressure test value 1	7	Pressure test value 1	12	Mean pressure value
3	Pressure test value 1	8	Pressure test value 1	13	Standard deviation value
4	Pressure test value 1	9	Pressure test value 1	—	—
5	Pressure test value 1	10	Pressure test value 1	—	—

Compared to subjective weighting approaches such as expert scoring or the analytic hierarchy process (AHP), CRITIC provides a data-driven framework that simultaneously considers the amount of information each indicator carries (via standard deviation) and the degree of redundancy between indicators (via intercriterion correlation coefficients).²⁹

The computational workflow for weight determination proceeds through three sequential steps, and the weight values are listed in Table 2.

Table 2.

Weight values of the eight anatomical test points

Test point	Position
Test point	1	2	3	4	5	6	7	8	9	10	11	12	13
Shoulder left	9.61%	4.96%	7.18%	4.33%	18.96%	9.81%	11.89%	4.16%	13.19%	4.86%	6.79%	3.22%	1.06%
Shoulder right	8.53%	12.91%	4.07%	8.23%	8.76%	6.96%	5.93%	7.76%	12.93%	10.20%	6.38%	3.87%	3.47%
Waist front	2.60%	11.52%	8.43%	7.27%	7.88%	7.05%	10.06%	12.94%	7.13%	9.05%	7.61%	5.61%	2.85%
Waist back	1.77%	7.01%	10.02%	5.71%	12.22%	14.18%	11.36%	4.25%	10.44%	10.07%	5.73%	4.42%	2.82%
Waist left	7.02%	5.27%	6.72%	7.17%	22.70%	9.19%	6.73%	0.77%	5.63%	13.56%	8.67%	3.27%	3.31%
Waist right	3.80%	8.99%	17.42%	21.53%	17.23%	2.46%	1.83%	4.41%	2.87%	5.38%	9.13%	2.81%	2.14%
Leg left	9.58%	8.46%	6.77%	13.79%	5.24%	10.13%	6.79%	14.37%	8.49%	2.86%	7.33%	2.54%	3.65%
Leg right	16.28%	8.30%	4.32%	5.27%	2.43%	8.94%	8.78%	7.01%	13.50%	6.01%	11.45%	2.91%	4.80%

Construct the initial evaluation matrix X, where n samples and m evaluation indicators are systematically organized as specified in

X = {[x_{i j} ​]}_{m \times n} = (\begin{matrix} x_{11} & \dots & x_{1 m} \\ ⋮ & ⋱ & ⋮ \\ x_{n 1} & \dots & x_{n m} \end{matrix})

(1)

For each standardized criterion j, the standard deviation (δ_j) is computed to quantify data dispersion, while the correlation coefficient (r_ij, where i, j = 1, 2, …, n) between criteria is calculated.

2. Compute the information content C_j for each evaluation index using

C_{j} = δ_{j} \sum_{i = 1}^{n} (1 - r_{i j})

(2)

where higher C_j values indicate greater informational significance of the jth indicator within the evaluation framework.³⁰

3. Determine the final weight W_j for each evaluation index through normalization of information content

W_{j} = \frac{C_{j}}{\sum_{j = 1}^{m} C_{j}}

(3)

where higher W_j values correspond to greater decision-making significance of the jth indicator.

Using the computed weights, the technique for order preference by similarity to ideal solution (TOPSIS) method was applied to compute the composite performance scores for each harness configuration across bilateral shoulder regions, anterior/posterior waist segments, and left/right thigh sections.

The TOPSIS method was selected as the core decision-making tool to evaluate harness configurations. Unlike qualitative ranking or subjective scoring systems, TOPSIS quantitatively measures the relative closeness of each design to an ideal solution across multiple performance indicators. This feature is suitable for our study, where trade-offs between safety (minimizing peak values) and comfort (achieving distribution uniformity) must be balanced. In addition, TOPSIS complements the CRITIC weighting system, which assigns weights based on data dispersion and intercriterion correlation, ensuring an unbiased and data-driven evaluation process.

Employing the standardized pressure sample matrix collected in this study, the weighted matrix (v_ij =W_j × r_ij) was calculated using the CRITIC-derived weights. The ideal solution A⁺ (minimum pressure) and the negative ideal solution A⁻ (maximum pressure) were determined by taking the column-wise minima and maxima of the weighted matrix, respectively³¹:

A^{+} = \min (v_{i j} ​), A^{-} = \max (v_{i j} ​)

(4)

The Euclidean distances to these reference points were computed as

D_{i}^{+} = \sqrt{\sum {(v_{i j} - A_{j}^{+})}^{2}}, D_{i}^{-} = \sqrt{\sum {(v_{i j} - A_{j}^{-})}^{2}}

(5)

The relative closeness coefficient C_i, serving as the ranking criterion, was then derived:

C_{i} ​ = \frac{D_{i}^{-}}{D_{i}^{+} ​ + D_{i}^{-}}

(6)

The optimal design configuration, based on these weighted rankings, is summarized in Table 3.

Table 3.

Optimal design scheme

Region	Score				Rank
Region	S1	S2	S3	S4	Rank
Shoulder left	0.82	0.67	0.46	0.25	S1 > S4 > S3 > S2
Shoulder right	0.44	0.21	0.43	0.55	S4 > S1 > S3 > S2
Waist front	0.81	0.59	0.49	0.24	S1 > S3 > S4 > S2
Waist back	0.24	0.49	0.60	0.64	S4 > S1 > S3 > S2
Waist left	0.65	0.53	0.37	0.54	S1 > S4 > S2 > S3
Waist right	0.46	0.33	0.57	0.62	S1 > S4 > S2 > S3
Leg left	0.28	0.17	0.55	0.59	S1 > S4 > S2 > S3
Leg right	0.80	0.76	0.47	0.64	S1 > S4 > S2 > S3

The final optimal design configuration was subsequently implemented in CLO3D, with the refined virtual prototype and its performance characteristics illustrated in Figure 4.

Figure 4.

Optimal solution model: (a) planar view; (b) structural view; (c) wearing view.

Specifically, for the shoulder region, the H-shaped strap was selected, as it demonstrated the lowest average and peak pressures compared with the X-, Y-, and V-shaped alternatives. For the waist region, a full-coverage waist strap was adopted due to its broader contact area and better pressure distribution. For the leg region, the parallel-shaped leg strap was chosen over the cross-shaped strap, which exhibited higher peak pressure values and less uniform pressure profiles.

Numerical simulation

To evaluate the mechanical behavior of the harness under realistic wearing conditions, a FEA simulation framework was developed.

Model establishment

The computed tomography (CT) scan images of a healthy male subject (age 23 years, height 178 cm, mass 74 kg, body mass index 23.8 kg/m²) were imported into Mimics18.0 software (Materialise Inc., Leuven, Belgium) for 3D reconstruction. Subsequently, the models were imported into Geomagic Studio 11 software (Geomagic, Inc., Research Triangle Park, NC, USA) for polygon processing and reconstruction into NURBS smooth surface models, thereby establishing personalized geometric models.

Furthermore, by importing anthropometric data derived from CT scans into the CLO3D wear simulation model and subsequently performing CT-based calibration of geometric modeling parameters and joint positions within the software, we ensured anatomical consistency of the resultant models. This methodological harmonization guarantees cross-platform simulation-result comparability.

The research was approved by the Shanghai Yangzhi Rehabilitation Hospital (SBKT-2022-008). Before the CT scan, informed consent has been obtained from the subject.

Simulation process

The human–harness pressure data simulated through CLO3D’s virtual fitting pressure mapping were imported into Abaqus/Standard version 6.14 (Simulia Inc., Providence, RI, USA) to simulate the interactive biomechanical process and assess the soft tissue stress–strain distribution in response to external loading conditions. As indicated by Table 4, the pressure values from eight critical measurement points were extracted.

Table 4.

Input pressure values

Position	Pressure value (kPa)	Position	Pressure value (kPa)
Shoulder left	6.453	Waist front	4.723
Shoulder right	5.716	Waist back	2.200
Leg left	2.625	Waist left	2.163
Leg right	2.320	Waist right	2.437

In the pressure simulation, the model analyzes pressure distribution and mechanical stresses under static loading conditions. While the simulation does not fully capture the dynamic complexity of forces during active movement, it effectively represents the major force application zones and their interactions with soft tissues. This approach allows us to identify potential high-pressure areas that may cause discomfort or soft tissue injury, providing valuable insights into harness design improvements.

FEA

The personalized 3D human model was analyzed in Abaqus using a structured FEA framework. Mesh discretization was performed in HyperMesh (version 11.0, Altair Engineering, Inc., Executive Park, CA, USA), generating a grid of 927,000 linear tetrahedral elements, with convergence studies confirming an optimal balance between computational efficiency and simulation accuracy.

The material properties of skin, fat, muscle, and bone tissues were defined based on validated constitutive laws derived from the literature. Skin was modeled as a linear elastic material, whereas both fat and muscle tissues were modeled as hyperelastic, nearly incompressible soft tissues using the first-order Ogden model. These parameters were derived from in vivo and ex vivo experimental studies reported by Hedrich et ak.,³² Hadid et al.,²⁰ and Fallahnezhad et al.³³ Bone was treated as a rigid body, given its minimal deformation under harness-related loading conditions (detailed parameters are given in Table 5).³⁴

Table 5.

Detailed material properties of our model

Property	Tissue
Property	Skin^20,32	Fat^20,33	Muscle^20,33	Bone³⁴
Material model	Elastic	Hyperelastic	Hyperelastic	Rigid
Property	E = 0.15 MPav = 0.46	μ_F = 1166.7 Pa,α_F = 16.2	μ_M = 1907.4 Pa,α_M = 4.6	—^a
Mass density (kg/m³)	1100	920	1100	1000
Modulus (MPa)	2.038	3.400	0.699	20,100
Poisson’s ratio	0.49	0.49	0.49	0.3

Parameters that are not required for certain material definitions.

Boundary conditions constrained lower extremities to simulate weight-bearing, whereas external loads were applied according to CLO3D-derived pressure distributions.

For each simulation, peak von Mises stress and principal shear strain values in the skin at eight critical anatomical measurement points were computationally extracted using MATLAB R2020a (The MathWorks Inc., Natick, Massachusetts, USA).

In addition, we explicitly acknowledge the limitations inherent in single-subject modeling and emphasize the necessity for future investigations to incorporate a broader spectrum of body morphologies, including but not limited to elderly individuals, pediatric populations, and patients with musculoskeletal conditions.

Harness pressure optimization

Wearer comfort during use remains underexplored, and harness designs seldom integrate biomechanical factors of the user.^35,36 To improve comfort and biomechanical performance, an optimization framework was developed to enhance pressure distribution. Given its advanced simulation automation and design optimization capabilities, iSIGHT (Dassault Systèmes Deutschland, Hamburg, Germany) was employed to execute the optimization process.

Problem definition

The optimization of human–harness interface pressure is formulated as a multiobjective problem integrating FEA with iSIGHT optimization techniques. This methodology iteratively adjusts input pressures at designated anatomical measurement points to minimize global stress–strain responses in the human body model. Eight critical anatomical regions were systematically evaluated: Shoulder Left Stress (S_L_S), Shoulder Right Stress (S_R_S), Shoulder Left Strain (S_L_E), Shoulder Right Strain (S_R_E), Waist Front Stress (W_F_S), Waist Back Stress (W_B_S), Waist Left Stress (W_L_S), Waist Right Stress (W_R_S), Waist Front Strain (W_F_E), Waist Back Strain (W_B_E), Waist Left Strain (W_L_E), Waist Right Strain (W_R_E), Leg Left Stress (L_L_S), Leg Right Stress (L_R_S), Leg Left Strain (L_L_E), and Leg Right Strain (L_R_E). Initial pressure parameters are derived from CLO3D simulation of the optimized harness configuration, whereas baseline stress–strain data were obtained through Abaqus computational modeling. The optimization objective focuses on determining the optimal pressure distribution across all measurement points to concurrently minimize biomechanical stress–strain responses and maintain biological tolerability thresholds.

iSIGHT model configuration

The Abaqus-based human body model was integrated into iSIGHT for automated optimization. Figure 5 illustrates the workflow implemented in iSIGHT.

Figure 5.

Workflow Task 1 in iSIGHT.

The optimization workflow adopted a two-stage strategy comprising design of experiments (DOE) and optimization algorithm implementation.³⁷ In the DOE phase, variable matrices were systematically sampled across predefined ranges to identify influential parameters, whereas preliminary analysis screened the design space and provided estimated optimal parameter sets. Subsequently, the optimization phase utilized DOE results as initial design references, followed by iterative refinement through global search algorithms.³⁸ During each iteration, FEA models were automatically executed, with iSIGHT algorithmic adjustments progressively refining the solution to enhance convergence efficiency.

Parameter input

To ensure the global exploration capability of the algorithm, a global exploration method from the iSIGHT platform was selected.³⁹ Specifically, adaptive simulated annealing (ASA) was chosen due to its robust global search capability and proven efficiency in navigating high-dimensional and nonlinear optimization problems for the multiobjective optimization phase. Compared with other algorithms (e.g., genetic algorithms and particle swarm optimization), ASA demonstrated superior convergence performance in our preliminary trials within the iSIGHT environment, especially in problems involving complex biomechanical constraints and pressure response feedback loops.⁴⁰

During mathematical model formulation, both rehabilitation harness comfort requirements and necessary support forces were holistically evaluated, with model optimization constrained within biologically tolerable pressure thresholds.⁴¹

The mathematical formulation is defined as

\min f (\begin{array}{l} L_L_E, L_L_S, L_R_E, L_R_S, S_L_E, S_L_S, \\ S_R_E, S_R_S, W_F_E, W_F_S, W_B_E, \\ W_B_S, W_L_E, W_L_S, W_R_E, W_R_S \end{array})

(7)

Equation (7) represents the objective function of the optimization process, which aims to minimize the overall stress and strain experienced across all relevant anatomical regions. This objective is central to improving harness comfort and reducing the risk of stress-related injuries.

To ensure both the clinical safety and comfort of the rehabilitation harness design, our optimization model incorporated multiobjective constraints based on physiologically validated thresholds for interface pressure and internal soft tissue deformation.^16–18,21 The constraints in the mathematical model are defined as follows:

\begin{array}{l} 0.0020 < L_L < 0.0133, 0.0020 < L_R < 0.0133, \\ 0.0059 < S_L < 0.0200, 0.0051 < S_R < 0.0200, \\ 0.0020 < W_B < 0.0133, 0.0020 < W_F < 0.0133, \\ 0.0020 < W_L < 0.0167, 0.0020 < W_R < 0.0167 \end{array}

(8)

\begin{array}{l} L_L_E, L_R_E, S_L_E, S_R_E, \\ W_F_E, W_B_E, W_L_E, W_R_E \leq 0.5 \\ L_L_S, L_R_S, S_L_S, S_R_S, \\ W_F_S, W_B_S, W_L_S, W_R_S \leq 6.0 \end{array}

(9)

The algorithm configuration established design variable optimization ranges equivalent to the initial pressure values extracted from CLO3D simulations. The ASA algorithm was implemented with explicitly defined parameters, while retaining default values for unspecified variables. Key parameter configurations included the number of designs for convergence check set to 5 and the maximum number of generated designs limited to 1000, ensuring systematic exploration of the design space.

Inverse calculation of optimal harness material parameters

Based on the optimal pressure values obtained through algorithmic optimization at various body locations, this study aims to inversely derive the optimal material parameters for the harness.

The workflow in iSIGHT for this task follows a process similar to that of the previous section, as depicted in Figure 6. Before applying the optimization algorithm for inverse calculation, a mathematical mapping model was first established to link the material parameters of the harness with the corresponding contact pressures.

Figure 6.

Workflow Task 2 in iSIGHT.

Construction of approximate contact pressure model and error analysis

Preliminary investigations integrating textile mechanical testing and experimental trials,⁴² identified fabric material as a critical determinant of wearing comfort, consequently establishing material parameters of the harness as pivotal considerations in human–harness interface design.

Based on their fundamental roles in governing human–device interface mechanics, this study selects warp strength (σ_T), weft strength (σ_W), density (ρ), and thickness (T) as core material parameters for optimization. Specifically, warp and weft strengths modulate stress distribution under biomechanical loading, directly influencing pressure concentration effects at the interface.^9,43 Density regulates interface friction, compressive resistance, and porosity-mediated mechanisms, with optimal ranges effectively mitigating shear-induced tissue damage.^44–48 Thickness governs mechanical compliance and contact area, critically determining load-spreading efficiency.^44,49 Through synergistic optimization of these material parameters in framework, clinically significant pressure distribution modulation has been achieved.

This study employed the optimal Latin hypercube design methodology to guide DOE, generating a sample matrix comprising 120 experimental points. This configuration ensures the constructed design matrix simultaneously satisfies the fundamental requirements for mathematical model fitting while achieving comprehensive coverage of the design space.^50,51 Pressure analysis was systematically performed using CLO3D software for each material parameter combination.

Based on systematic analysis of harness–body contact areas and the distribution characteristics of eight pressure measurement points, while considering anatomical symmetry to ensure balanced representation, adjacent measurement regions with similar pressure distribution characteristics and optimization objectives were consolidated. The original eight pressure measurement points were thus merged into four key regions: shoulder pressure (S), lumbar pressure (W₁), anterior abdominal pressure (W₂), and leg pressure (L).

Pressure measurements were recorded for each test group across the four anatomical regions. Mean pressure values were computed for each parameter combination, followed by comprehensive error analysis through multiple sampling iterations.

The linear regression, ridge regression, hierarchical regression, least absolute shrinkage and selection operator (LASSO) regression, and partial least squares regression (PLSR) were evaluated in MATLAB R2020a. Models were selected based on optimal R² values approaching unity, with linear regression demonstrating superior performance across all regions. The comparative fitting results are presented in Figure 7.

Figure 7.

Comparison between the real value and the predicted value.

Variance inflation factor (VIF) values remained below 10, confirming the absence of multicollinearity issues and validating model integrity. In addition, while the regional consolidation during function fitting may obscure individual pressure peaks, the robustness of our dataset ensures the validity of the fitted equations. The high R² values and acceptable VIF levels confirm that the derived formulas meet the accuracy requirements for simulation purposes.

The derived formulas for the four regions (shoulder, lumbar, abdomen, and leg) were derived as follows (units for S, L, W₁, and W₂ are kPa):

\begin{array}{l} S = 0.517 + 2.265 \times 10^{- 5} σ_{T} + 2.495 \times 10^{- 5} σ_{W} \\ + 0.0022 ρ - 0.005 T \end{array}

(10)

\begin{array}{l} W_{1} = - 0.572 + 9.953 \times 10^{- 6} σ_{T} + 1.752 \times 10^{- 5} σ_{W} \\ + 0.0002 ρ + 0.110 T \end{array}

(11)

\begin{array}{l} W_{2} = - 0.090 + 2.060 \times 10^{- 6} σ_{T} + 3.383 \times 10^{- 5} σ_{W} \\ - 4.693 \times 10^{- 5} ρ + 0.245 T \end{array}

(12)

\begin{array}{l} L = - 1.625 + 2.637 \times 10^{- 6} σ_{T} + 5.295 \times 10^{- 6} σ_{W} \\ + 0.0004 ρ + 0.270 T \end{array}

(13)

Inversion algorithm selection

The optimization framework in iSIGHT was employed to solve for the material parameters using the derived regression models. In this approach, material parameters (σ_T, σ_W, ρ, T) were configured as variables, and regional pressures (S, L, W₁, W₂) were treated as targets based on the approximation models.

The ASA algorithm was selected due to its efficiency in global search and optimization, with parameter configurations consistent with those described in the preceding section. The objective function was defined based on the optimal contact pressures obtained from the previous optimization step, with target values as follows:

Target (S = 5.476, L = 3.126, W_{1} = 2.475, W_{2} = 4.255)

(14)

Results and discussion

Human–harness interface pressure optimization

The optimal pressure distribution, derived from minimizing the stress–strain values across each body part through systematic optimization, is presented in Figure 8.

Figure 8.

Optimal pressure distribution.

To validate the optimization efficacy, the verification protocol involved reimporting the optimized interface pressure parameters into Abaqus for comparative analysis of stress–strain characteristics at eight anatomical landmarks before and after optimization. The results demonstrated an approximate 15% reduction in stress–strain values at critical body regions. The verification results are shown in Figure 9.

Figure 9.

Verification of interface pressure.

As illustrated in Figure 9, the optimized pressure distribution achieved a 15.47% reduction in maximum stress (W_F_S: from 25.3 to 21.4 kPa) and 19.58% strain reduction (S_R_E: from 0.12 to 0.096), respectively, compared with the baseline harness design. Notably, stress and strain in critical regions such as the shoulders (S_L_S: decrease 14.70%; S_R_S: decrease 10.95%) and legs (L_L_S: decrease 9.92%; L_R_S: decrease 9.96%) exhibited reductions of approximately 10%, consistent with clinical reports highlighting these areas as high-risk zones for stress injuries.^36,52,53

It is important to note that while the optimized pressure distributions led to reductions in peak von Mises stress and shear strain within simulated soft tissues, the damage thresholds referenced are based on computational modeling studies, not on direct in vivo measurements. As such, the physiological implications of these findings should be interpreted with caution and warrant further experimental validation.

To achieve the predefined objective function of minimizing localized stress–strain distributions, the pressure of the harness was strategically redistributed during the optimization process. This approach of redistributing pressure is a common methodology for improving wearer comfort and optimizing harness design.⁵³ Specifically, pressure on the left and right waist regions was increased, whereas reductions were implemented for the shoulders, legs, and front/back waist regions. These reductions are clinically significant, as they bring stress and strain values below established thresholds for tissue damage. Furthermore, according to multiple studies, the reduction in these values has been shown to effectively mitigate wearer injuries and enhance wearing comfort.^54–56

These findings underscore the necessity of reinforcing support in the lumbar regions while alleviating pressure loads on the shoulders and legs. For instance, strain in the left leg harness was carefully balanced against overall load reduction, ensuring localized reinforcement without compromising global comfort. This design strategy not only optimizes the stress–strain distribution across the body during rehabilitation but also reduces injury risks by 15–20% in high-incidence areas. Consequently, the proposed adjustments enhance both biomechanical safety and user comfort, addressing key challenges in clinical rehabilitation protocols.

However, the biomechanical tolerance is not uniform across populations and loading conditions. For instance, elderly individuals typically exhibit lower tissue elasticity, thinner subcutaneous fat layers, and reduced musculoskeletal compensation capacity, which makes them more vulnerable to stress-related injuries.^57,58 Dynamic loading associated with gait cycles, particularly in body-weight-supported rehabilitation scenarios, could introduce additional stress peaks and cyclic fatigue effects. While our simulations are based on the anthropometric and tissue parameters of an average healthy adult male under static loading conditions, the modeling framework can be extended in future studies by incorporating population-specific parametric human models and including dynamic loading profiles.

Harness material parameter optimization

A comparison of the optimal pressure distribution reveals significant differences in stress variations across the shoulders, waist, and legs. Under the condition of modifying only the material parameters while maintaining the harness tightness during use, the harness materials for different regions were optimized based on the structural design illustrated in Figure 4. The corresponding optimization results are presented in Table 6.

Table 6.

Optimization values of material parameters

Position	Material parameter	Initial value	Optimization value
Shoulder	Warp strength (g/s)	100,000	90,231.649
	Weft strength (g/s²)	100,000	89,516.961
	Thickness (mm)	3	2.227
	Density (g/m²)	300	287.605
Lumbar region	Warp strength (g/s²)	100,000	112,886.640
	Weft strength (g/s²)	100,000	62,898.097
	Thickness (mm)	3	3.869
	Density (g/m²)	300	209.190
Anterior abdominal region	Warp strength (g/s²)	100,000	136,050.420
	Weft strength (g/s²)	100,000	70,599.585
	Thickness (mm)	3	3.604
	Density (g/m²)	300	299.833
Leg	Warp strength (g/s²)	100,000	135,539.674
	Weft strength (g/s²)	100,000	95,670.866
	Thickness (mm)	3	3.430
	Density (g/m²)	300	265.589

Figure 10 illustrates the correlations among optimization parameters, delineating the interactions between input variables, σ_T, σ_W, ρ, and T, and output pressure parameters, including shoulder pressure (S), leg pressure (L), lumbar region pressure (W₁), and anterior abdominal region pressure (W₂). The analysis further quantifies the influence of each input parameter on the pressure distribution. The results demonstrate that weft strength exerts the most significant impact on pressure magnitudes across all regions, with shoulder, leg, lumbar, and anterior abdominal pressures exhibiting pronounced sensitivity to σ_W variations.⁵⁹ In contrast, material thickness demonstrates a comparatively marginal contribution to pressure modulation. These findings underscore the necessity of prioritizing weft strength control in the design and optimization of material parameter to achieve targeted mechanical performance, a principle extensively documented in textile research for diverse fabric systems.^60,61

Figure 10.

Correlation table.

The optimized parameters for each anatomical region, as detailed in Table 6, reveal distinct trends. For the lumbar region, the warp strength increased by 12.89%, whereas the weft strength decreased by 37.10%. This inverse relationship highlights the critical role of enhanced warp and weft strength in improving stiffness, which effectively redistributes stress concentrations and augments mechanical support in the lumbar area.^35,61,62 Conversely, the anterior abdominal region exhibited a 36.05% increase in warp strength alongside a 29.40% reduction in weft strength, suggesting a design strategy that balances localized rigidity with flexibility.

Nonetheless, it is essential to emphasize that the observed biomechanical advantages of increased lumbar support do not imply a one-size-fits-all recommendation. These findings should be regarded as a theoretical guideline for calibrated lumbar stiffness enhancement, rather than a directive for aggressive structural reinforcement. Such modifications should be implemented with individualized fitting strategies, material flexibility considerations, and clinical oversight.

The shoulder region exhibited concurrent reductions in both warp and weft strengths, with quantitatively measured decreases of 9.77% in warp strength and 10.48% in weft strength. Meanwhile, in the leg harness design, the warp strength was enhanced to reinforce structural support along the principal loading direction of gait motion, thereby improving tensile integrity and preventing material elongation or displacement during use, while the weft strength governing skin-contact compliance was strategically reduced by 4.33% to enhance surface conformability and interface comfort. This differential adjustment strategy enables the harness to maintain mechanical stability under tension while preserving soft tissue interface properties, achieving an optimal balance between biomechanical support and wearable comfort through direction-specific material optimization.¹⁹

Regarding the feasibility of the optimized material parameters, significant improvements in tensile strength can be achieved through multiple strategies: (i) densifying warp yarn density by increasing ends per inch (EPI); (ii) employing high-denier continuous filaments (e.g., 1680D–2000D); (iii) hybridizing yarns with high-tenacity fibers (e.g., PET/nylon or aramid blends); and (iv) incorporating finishing techniques such as resin coating or calendaring.

Material density and thickness further contribute to this differentiated design approach. For instance, the lumbar region’s density decreased by 30.27%, whereas its thickness increased by 28.97% (to 3.869 mm), enhancing compliance without compromising structural integrity. Similarly, the anterior abdominal region’s near-optimal density retention and moderate thickness reduction (3.604 mm) reflect a balance between flexibility and load-bearing capacity, aligning with wearer comfort requirements for such harness systems.³⁶

To validate the optimization effectiveness, the optimized material parameters were imported into CLO3D software for computational validation of pressure values. A comparative visualization of the optimized pressure distribution versus the simulated validation results is presented in Figure 11. The results demonstrate a consistent pressure reduction trend in major load-bearing regions (e.g., shoulders and lateral waist), with pressure decreases exceeding 25% in several structural configurations, achieving the anticipated pressure distribution profile.

Figure 11.

Verification of material parameters.

In summary, the data-driven optimization framework advocates for region-specific material strategies: prioritizing high-stiffness, warp-dominant configurations for the lumbar region to dissipate stress, and adopting reduced-strength, softer materials for the shoulder and leg regions to alleviate pressure. Such tailored approaches harmonize mechanical distribution at the human–harness interface, advancing both safety and comfort in rehabilitation applications.

Conclusion

This study developed a digital design framework by integrating cross-domain modeling and simulation tools to optimize harness structure and material parameters. The framework enabled systematic optimization of contact pressure and material properties (warp strength, weft strength, density, and thickness), resulting in a 15.47% reduction in von Mises stress and a 19.58% reduction in shear strain compared with conventional designs. These reductions are significant, as they bring stress and strain values below thresholds associated with tissue damage. The optimized design emphasizes region-specific material strategies: stiffer, high-warp-strength materials for the lumbar region to enhance load redistribution and softer materials with reduced warp and weft strength for the shoulders and legs to minimize pressure concentrations.

The evaluation of harness mechanical performance was conducted under static conditions using a FEA framework. Although this method offers precise control over boundary conditions, material parameters, and anatomical modeling, it does not fully replicate the complex, dynamic biomechanical interactions occurring during real-world rehabilitation activities such as gait, balance recovery, or postural transitions. Future work should incorporate experimental validation through human-subject testing.

Meanwhile, the study has limitations, including the exclusion of advanced fabric characteristics (e.g., weave patterns, anisotropic behaviors) and dynamic biomechanical interactions during movement. In addition, physiological comfort metrics (e.g., skin microclimate, user-reported comfort) were not validated. Future work should focus on investigating multiparameter coupling effects, conducting experimental validations through in vivo trials, and integrating machine-learning-driven surrogate models to improve predictive capabilities and extend the framework to applications in wearable robotics and rehabilitation devices.

To effectively support technology translation, this study outlines two key future research directions: (1) user trials involving controlled experiments with both healthy and rehabilitation subjects to systematically evaluate harness comfort, usability, and biomechanical responses during assisted gait; (2) machine learning integration to develop adaptive pressure adjustment strategies utilizing real-time feedback mechanisms, enabling personalized pressure distribution optimization based on individual anatomical characteristics and activity patterns.

Footnotes

Ethics approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration. The research was approved by the Shanghai Yangzhi Rehabilitation Hospital Department of Ethics Committee (SBKT-2022-008). Informed consent was obtained from all individual participants included in the study. Additional informed consent was obtained from all individual participants for whom identifying information is included in this article.

Data availability

The data generated or analyzed during this study are available from the corresponding author upon reasonable request. The code generated or used during the study is also available from the corresponding author upon reasonable request.

Declaration of conflicting of interests

The authors have no competing interests to declare that are relevant to the content of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the National Natural Science Foundation of China (grant number 12302423), the Shanghai Yangzhi Rehabilitation Hospital (Shanghai Sunshine Rehabilitation Centre) Sunshine Clinical Research Incubation Project (grant number 2024CRZH009), and the Wuhan Textile and Apparel Digital Engineering Technology Research Center Foundation (grant number SZH202401).

ORCID iDs

Yizhuo Chen

Jing Gao

Shitan Wang

References

Powell

Lopez

Westgate

Hines

Sawaki

Effects of dynamic overground body weight support training during inpatient rehabilitation after traumatic spinal cord injury: A retrospective case series. Amer J Phys Med Rehab 2022; 101(2): 196–200. DOI: 10.1097/phm.0000000000001828.

Bannwart

Bayer

Ignasiak

Bolliger

Rauter

Easthope

CA.

Mediolateral damping of an overhead body weight support system assists stability during treadmill walking. J Neuroeng Rehab 2020; 17(1): 108. DOI: 10.1186/s12984-020-00735-w.

Ettema

Pennink

Buurke

TJW

David

van Bennekom

CAM

Houdijk

Clinical indications and protocol considerations for selecting initial body weight support levels in gait rehabilitation: A systematic review. J Neuroeng Rehab 2024; 21(1): 97. DOI: 10.1186/s12984-024-01389-8.

Hadid

Gozes

Atoon

Gefen

Epstein

Effects of an improved biomechanical backpack strap design on load transfer to the shoulder soft tissues. J Biomech 2018; 76: 45–52. DOI: 10.1016/j.jbiomech.2018.05.016.

Borggraefe

Klaiber

Schuler

, et al. Safety of robotic-assisted treadmill therapy in children and adolescents with gait impairment: A bi-centre survey. Development Neurorehab 2010; 13(2): 114–119. DOI: 10.3109/17518420903321767.

Edsberg

Black

Goldberg

McNichol

Moore

Sieggreen

Revised national pressure ulcer advisory panel pressure injury staging system revised pressure injury staging system. J Wound Ostomy Continence Nurs 2016; 43(6): 585–597. DOI: 10.1097/won.0000000000000281.

Plooij

Apte

Keller

, et al. Neglected physical human–robot interaction may explain variable outcomes in gait neurorehabilitation research. Science Robotics 2021; 6(58): eabf1888. DOI: 10.1126/scirobotics.abf1888.

Varghese

Thilagavathi

Handle, fit and pressure comfort of silk/hybrid yarn woven stretch fabrics. Fibers Polym 2016; 17(3): 484–494. DOI: 10.1007/s12221-016-5561-5.

Kanai

Ogawa

Sasagawa

Shibata

Influence of the mechanical stretch property of fabrics on the wear comfort of men’s suit pants. Text Res J 2021; 91(23–24): 2771–2785. DOI: 10.1177/00405175211019131.

10.

Varghese

Thilagavathi

Development of woven stretch fabrics and analysis on handle, stretch, and pressure comfort. J Text Inst 2015; 106(3): 242–252. DOI: 10.1080/00405000.2014.914652.

11.

Tian

Evaluation of dress fit based on CLO3D virtual fitting technique. J Donghua Univ Nat Sci Ed 2018; 44(3): 397–402.

12.

Song

Research on the dressing effect evaluation of Qipao based on CLO3D virtual fitting technology. J Silk 2021; 58(12): 73–79.

13.

Youn

Zhan

Mathur

Advanced virtual fit technology for precision pressure application in medical compression waistbands. Appl Sci Basel 2024; 14(22): 10697. DOI: 10.3390/app142210697.

14.

Tian

Yang

Zhang

Sun

Shi

Chen

YH.

Design optimization of hydraulic machinery based on iSIGHT software: A review of methods and applications. Water 2023; 15(11): 2100. DOI: 10.3390/w15112100.

15.

Baszczynski

Effects of safety harnesses protecting against falls from a height on the user’s body in suspension. Int J Environ Res Public Health 2023; 20(1): 71. DOI: 10.3390/ijerph20010071.

16.

Ciaccia

Sznelwar

LI.

An approach to aircraft seat comfort using interface pressure mapping. Work 2012; 41: 240–245. DOI: 10.3233/wor-2012-0163-240.

17.

Mitsuno

Kai

Distribution of the preferred clothing pressure over the whole body. Text Res J 2019; 89(11): 2187–2198. DOI: 10.1177/0040517518786272.

18.

Kermavnar

O’Sullivan

Casey

de Eyto

O’Sullivan

LW.

Circumferential tissue compression at the lower limb during walking, and its effect on discomfort, pain and tissue oxygenation: Application to soft exoskeleton design. Appl Ergonom 2020; 86: 103093. DOI: 10.1016/j.apergo.2020.103093.

19.

Tadesse

Loghin

Dulgheriu

Loghin

Comfort evaluation of wearable functional textiles. Materials 2021; 14(21): 6466. DOI: 10.3390/ma14216466.

20.

Hadid

Belzer

Shabshin

Zeilig

Gefen

Epstein

The effect of mechanical strains in soft tissues of the shoulder during load carriage. J Biomech 2015; 48(15): 4160–4165. DOI: 10.1016/j.jbiomech.2015.10.020.

21.

Luboz

Bailet

Grivot

, et al. Personalized modeling for real-time pressure ulcer prevention in sitting posture. J Tissue Viabil 2018; 27(1): 54–58. DOI: 10.1016/j.jtv.2017.06.002.

22.

Huang

CLO3D-based 3D virtual fitting technology of down jacket and simulation research on dynamic effect of cloth. Wireless Commun Mobile Comput 2022; 2022: 5835026. DOI: 10.1155/2022/5835026.

23.

Takai

Teramae

Noda

, et al. Development of split-force-controlled body weight support (SF-BWS) robot for gait rehabilitation. Front Human Neurosci 2023; 17: 1197380. DOI: 10.3389/fnhum.2023.1197380.

24.

Lung

Yang

Liau

Cheung

Jain

Jan

YK.

Dynamic changes in seating pressure gradient in wheelchair users with spinal cord injury. Assistive Technol 2020; 32(5): 277–286. DOI: 10.1080/10400435.2018.1546781.

25.

Yang

Guo

Huang

Zhang

CQ.

Biomechanical response of lumbar facet joints under follower preload: A finite element study. BMC Musculoskel Disord 2016; 17. DOI: 10.1186/s12891-016-0980-4.

26.

McGill

Kavcic

Harvey

Sitting on a chair or an exercise ball: Various perspectives to guide decision making. Clin Biomech 2006; 21(4): 353–360. DOI: 10.1016/j.clinbiomech.2005.11.006.

27.

Rakheja

Boileau

PE.

Study of human-seat interface pressure distribution under vertical vibration. Int J Ind Ergonom 1998; 21(6): 433–449. DOI: 10.1016/s0169-8141(96)00084-4.

28.

Hirao

Naito

Yamazaki

Pressure sensitivity of buttock and thigh as a key factor for understanding of sitting comfort. Appl Sci Basel 2022; 12(15): 7363. DOI: 10.3390/app12157363.

29.

Zhang

Fan

Gao

CB.

CRITID: Enhancing CRITIC with advanced independence testing for robust multi-criteria decision-making. Sci Rep 2024; 14(1): 25094. DOI: 10.1038/s41598-024-75992-z.

30.

Zhao

Wang

Xue

Yang

A construction schedule robustness measure based on improved prospect theory and the Copula-CRITIC method. Appl Sci Basel 2020; 10(6): 2013. DOI: 10.3390/app10062013.

31.

Yang

Duan

Long

XW.

A study on the application of coordinated TOPSIS in evaluation of robotics academic journals. Math Prob Eng 2018; 2018: 5456064. DOI: 10.1155/2018/5456064.

32.

Hedrich

Morbach

Reiser

Girschick

HJ.

New insights into adult and paediatric chronic non-bacterial osteomyelitis CNO. Curr Rheumatol Rep 2020; 22(9): 52. DOI: 10.1007/s11926-020-00928-1.

33.

Fallahnezhad

O’Rourke

Bahl

Thewlis

Taylor

The role of muscle forces and gait cycle discretization when assessing acetabular cup primary stability: A finite element study. Comput Meth Programs Biomed 2023; 230: 107351. DOI: 10.1016/j.cmpb.2023.107351.

34.

Bayraktar

Morgan

Niebur

Morris

Wong

Keaveny

TM.

Comparison of the elastic and yield properties of human femoral trabecular and cortical bone tissue. J Biomech 2004; 37(1): 27–35. DOI: 10.1016/s0021-9290(03)00257-4.

35.

Filgueiras

Rebelo

(eds.). Evaluation of concepts for lumbar support in harness. In: 6th International Symposium on Occupational Safety and Hygiene (SHO 2010), Guimaraes, Portugal, 2010, vol. 2, pp. 11–12.

36.

Sepehr

Jabbari

Naeini

, et al. Design and psychometric analysis of the safety harness usability and comfort assessment tool (SH UCAT). Heliyon 2023; 9(2): e13524. DOI: 10.1016/j.heliyon.2023.e13524.

37.

Huang

Zheng

Mei

Design and optimization method of a two-disk rotor system. Int J Turbo Jet-Engines 2016; 33(1): 1–8. DOI: 10.1515/tjj-2014-0033.

38.

Kluess

Mittelmeier

van Rienen

Bader

An automatic approach for calibrating dielectric bone properties by combining finite-element and optimization software tools. Comput Meth Biomech Biomed Eng 2016; 19(12): 1306–1313. DOI: 10.1080/10255842.2015.1131980.

39.

Shen

Wang

Chen

Feng

ML.

Optimization design and nonlinear bending of bio-inspired helicoidal composite laminated plates. Materials 2023; 16(13): 4550. DOI: 10.3390/ma16134550.

40.

Shi

Lin

Shi

GL.

Noninvasive reconstruction of internal heat source in biological tissue using adaptive simulated annealing algorithm. Sci Rep 2024; 14(1): 16379. DOI: 10.1038/s41598-024-67253-w.

41.

Nam

Yang

Kim

Koo

Song

Kim

Quantification of comfort for the development of binding parts in a standing rehabilitation robot. Sensors 2023; 23(4): 2206. DOI: 10.3390/s23042206.

42.

Wang

YY.

Effects of clothing ease and body postures on the air gap and clothing coverage. Int J Cloth Sci Technol 2019; 31(4): 578–594. DOI: 10.1108/ijcst-12-2018-0158.

43.

Mousazadegan

Ezazshahabi

Contribution of constituent yarns of the worsted woven fabrics in various directions during the tensile loading. J Eng Fibers Fabrics 2019; 14: 1–10. DOI: 10.1177/1558925019846694.

44.

Baby

Mathur

DenHartog

nondestructive quantitative evaluation of yarns and fabrics and determination of contact area of fabrics using the x-ray microcomputed tomography system for skin-textile friction analysis. ACS Appl Mater Interf 2021; 13(3): 4652–4664. DOI: 10.1021/acsami.0c18300.

45.

Derler

Rao

Ballistreri

Huber

Scheel-Sailer

Rossi

RM.

Medical textiles with low friction for decubitus prevention. Tribol Int 2012; 46(1): 208–214. DOI: 10.1016/j.triboint.2011.03.011.

46.

Rotaru

Pille

Lehmeier

, et al. Friction between human skin and medical textiles for decubitus prevention. Tribol Int 2013; 65: 91–96. DOI: 10.1016/j.triboint.2013.02.005.

47.

Rout

Singh

MK.

Porosity determination of textile fabrics: A novel mathematical approach and experimental validation. Mater Today Commun 2023; 37. DOI: 10.1016/j.mtcomm.2023.107559.

48.

Afroz

Rashid

Islam

Akter

Investigation on the effect of cotton-tencel (50:50) siro yarn twist multipliers on fabric stiffness. Heliyon 2022; 8(12): e12498. DOI: 10.1016/j.heliyon.2022.e12498.

49.

Stoffberg

Hunter

Botha

The effect of fabric structural parameters and fiber type on the comfort-related properties of commercial apparel fabrics. J Nat Fibers 2015; 12(6): 505–517. DOI: 10.1080/15440478.2014.967370.

50.

Wang

Nakashima

Lai

, et al. Modified algorithms for fast construction of optimal Latin-hypercube design. IEEE Access 2020; 8: 191644–191658. DOI: 10.1109/access.2020.3032122.

51.

Wang

Liu

Wei

XX.

Structure design method of new balanced vibration reduction gear for the three cylinder engine. PLoS ONE 2022; 17(4): e0266560. DOI: 10.1371/journal.pone.0266560.

52.

Gregory

Milosavljevic

Pal

Callaghan

JP.

An examination of shoulder kinematics and kinetics when using a commercial trunk harness while sheep shearing. Appl Ergonom 2008; 39(1): 29–35. DOI: 10.1016/j.apergo.2007.04.001.

53.

Nam

Kwon

Kim

Ahn

Development of pant-type harness with fabric air-pocket for pain relief. Appl Sci Basel 2019; 9(9): 1921. DOI: 10.3390/app9091921.

54.

Hsiao

Turner

Whisler

Zwiener

Impact of harness fit on suspension tolerance. Human Factors 2012; 54(3): 346–357. DOI: 10.1177/0018720811434962.

55.

Turner

Wassell

Whisler

Zwiener

Suspension tolerance in a full-body safety harness, and a prototype harness accessory. J Occup Environ Hygiene 2008; 5(4): 227–231. DOI: 10.1080/15459620801894386.

56.

Kalkhoven

Watsford

Impellizzeri

FM.

A conceptual model and detailed framework for stress-related, strain-related, and overuse athletic injury. J Sci Med Sport 2020; 23(8): 726–734. DOI: 10.1016/j.jsams.2020.02.002.

57.

Liu

Zhao

Zhang

, et al. Biomechanical assessment of different transforaminal lumbar interbody fusion constructs in normal and osteoporotic condition: A finite element analysis. Spine J 2024; 24(6): 1121–1131. DOI: 10.1016/j.spinee.2024.01.017.

58.

Huang

Long

(eds.). A study on the effect of age on the injury criteria and modeling of lower extremity long bone. In: 8th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA), Macau, PR China, 11–12 March 2016.

59.

Liu

Deng

Chen

(eds.). The mechanical properties study of woven fabrics. In: 2nd International Conference on Advanced Measurement and Test (AMT 2011), Nanchang, PR China, 2011, vol. 6, pp. 24–26.

60.

Xie

Liu

WB.

Effect of structural parameters on tensile properties of alumina fabrics. Text Res J 2023; 93(9–10): 2095–2102. DOI: 10.1177/00405175221136295.

61.

Baszczynski

New test method with a hybrid III anthropomorphic dummy for textile safety harnesses. Fibres Text East Eur 2020; 28(1): 81–86. DOI: 10.5604/01.3001.0013.5861.

62.

Fok

Yip

Yick

SP.

Design and fabrication of anisotropic textile brace for exerting corrective forces on spinal curvature. J Ind Text 2022; 51(1 Suppl): 1682S–1702S. DOI: 10.1177/15280837211032619.