EVALUATION OF A SURFACE SHAPE OPTIMIZATION TECHNIQUE FOR CUSTOM CONTOURED CUSHION DESIGN

Tricia E Karg , David M Brienza , Kao-Chi Chung  , Clifford E Brubaker ,  School of Health and Rehabilitation Sciences, University of Pittsburgh   Institute of Biomedical Engineering, National Cheng University, Taiwan

ABSTRACT

A system for the design and analysis of seat support and buttock tissue interfaces has been developed. The system has the ability to control the seating surface shape while measuring the external pressure applied to the buttocks by the surface. The system, therefore, facilitates the study of the relationships between support surface shape and interface pressure, and support surface shape and soft tissue distortion. Closed-loop control of the system allows for the dynamic formulation of a support surface on the basis of programmable criteria. The design, instrumentation and system performance test results are presented in a previous paper. This paper presents the basis for and evaluation of a support surface-shape optimization algorithm that has been developed using criteria for the measured parameters of interface pressure and tissue force-deflection characteristics. The technique has been tested on ten able-bodied subjects and 30 elderly subjects. Results presented indicate that resultant optimal support shapes carved into high resiliency foam cushions satisfy optimization criteria, provide lower peak interface pressures than flat foam and the support shape derived from the seated contour only, and reduce the tissue stiffness characteristic.

INTRODUCTION

The computer-aided seating system (CASS) is a dynamically controlled shape and pressure sensing system developed for quantification of the complex relationship between support surface shape, tissue thickness changes and interface pressures. The system consists of an instrumented seat support surface that measures interface pressure while controlling support surface shape, an interface unit for processing the pressure transducer signals and control of the drive motor array, and a DOS compatible personal computer for high level control of the system. The adjustable seat forms a 3-dimensional support surface through selective adjustment to the heights of support elements arranged in an 11 by 12 array. The 43 x 47 cm array of support elements can be adjusted vertically within a range of, approximately, 15 cm. Pressure sensors are fixed into swiveling heads on top of each support element. The top of the support element rotates freely so that the pressure transducers are oriented in a direction normal to direction of net force. The supporting seat structure was designed for flexibility so that the support surface could be used in a simulated seating environment. The structure includes an adjustable sling backrest, armrests, and a footrest. Seat depth, seat to back angle, armrest lateral position, armrest vertical position, footrest height and footrest angle are all adjustable. The system has shown to be a precise, repeatable, and reliable tool.

Software has been developed to control the motor array, communicate with the data acquisition processor to receive pressure measurements, and display support surface shape and pressure. This software is expandable to allow for the development of control algorithms using support surface shape and interface pressure as parameters. At present, an algorithm based on interface pressure and tissue thickness changes is used to drive the system to an "optimized" support surface shape. Tissue thickness changes are inferred from the force-deflection characteristics of the tissue as measured by external pressure per unit change in deflection, i.e., stiffness. The measurement of stiffness is used as an indictor of tissue thickness.

METHODS

This evaluation of the support surface optimization technique is the first to utilize the capabilities of the CASS. In this study an iterative algorithm that optimizes support surface shape using pressure measurements is used. The resulting optimal support surface shape is characterized by a pressure distribution that minimizes the inferred distortion of the buttocks soft tissues from their unloaded shape. The theoretical development of this algorithm is presented in detail in Brienza et al. [1].

Although the CASS system cannot measure unloaded shape or tissue thickness changes relative to thickness in the unloaded condition, this information is inferred from the measurements of tissue stiffness. Tissue stiffness is measured as a change in external pressure per unit change in deflection. Relative soft tissue thickness can be inferred from these stiffness measurements because the load deflection characteristic of soft tissue is such that thinner sections of loaded tissue will appear stiffer than thicker sections under identical loads. Using this property, the algorithm minimizes a quadratic performance index that is a measure of the summation of the gradient of the scaler quantity, stiffness multiplied by pressure, over the array of support elements. The successful minimization of the performance index results in a buttock and support surface interface condition for which the variations in relative distortion of the buttocks soft tissue is minimized. Equivalently a support surface is found for which there is an inverse relationship between the measured external pressure and tissue stiffness. In other words, areas of the support surface in contact with stiffer tissue have proportionately lower external pressures than areas in contact with more compliant tissues.

A pilot study was performed on 10 able-bodied persons between the ages of 18 and 40 to demonstrate the CASS performance and to obtain preliminary data on the tissue stiffness based control algorithm. A subsequent study was performed on 30 elderly subjects over the age of 65. The protocol for the study was as follows. The subject first sits on a mechanical seating system with passive spring loaded support elements that deflect under the load of the subjects weight to form the contoured support surface [2]. With assistance from the research team, the subject was positioned in a prescribed seated posture. Once positioned on the mechanical contour measurement device, the resulting shape of the support surface as measured by the deflections on the spring-loaded support elements is stored in the computer's memory. The subject then transfers from the mechanical measurement device to the CASS with the support elements forming a flat surface. Care is taken to ensure that the subject re-establishes the prescribed reproducible seated posture established on the mechanical measurement device. The iterative optimization algorithm is initiated by recording pressure and stiffness measurements with the support surface configured in the flat condition. After these initial measurements, the support surface elements are moved to form the initial support surface shape previously recorded on the mechanical system. Again, pressure and stiffness measurements are recorded and the support surface shape is adjusted toward the optimum shape as determined by the stiffness and pressure criteria. For each iteration, an optimum pressure distribution is determined from the pressure and stiffness measurements. The differences between the actual and optimum pressures are used to determine the adjustments needed to move toward the optimal support surface shape. After each adjustment, new pressure measurements are compared to optimal pressure values dependent on the optimization criteria. If the measured values are sufficiently close to the optimal values the algorithm has converged; if not, the computer determines the necessary adjustment in support surface shape and another iteration begins.

After determining the optimal support surface shape, the initial and optimal shapes are transferred to a cushion cutting machine that carves the shapes into high resiliency foam cushions. The efficacy of the cushions was evaluated by measuring the distribution of external pressure exerted on the buttocks while seated in the contoured cushion. This was done using a mat containing a 15 by 15 array of resistive sensors that measure pressure (FSA Pad, Vistamed, Manitoba, Canada.) The subject sat on three surfaces of flat foam, the initial contour, and the optimized contour with the FSA mat between the cushion and his or her buttocks.

RESULTS

Figure 1 shows the normalized pressure and stiffness distribution and the quantity of normalized pressure multiplied by normalized stiffness for a test subject. The figure shows these parameters from a single transverse row of support elements passing near to the location of the subjectÕs ischial tuberosities. Data is shown for the first iteration on a flat support surface (iteration 1) and then for selected iterations as the shape is altered by the optimization algorithm until convergence (iteration 10).

[INSERT] Figure 1. Characteristics of Shape, Stiffness, and Pressure through the Optimization Process

The results of the pilot study are summarized in Tables 1 and 2. Table 1 shows the maximum values of pressure, stiffness and deflection for three support surface shapes: flat, initial contour, and optimal contour, as measured by the CASS during the optimization procedure. Table 2 shows the results of the measurements taken with the pressure mat with the subject seated on the three corresponding foam surfaces. Data was available for only nine of the ten able-bodied research subjects. The external pressure distribution was measured for the three support surface shapes for an initial measurement and a measurement when the subject was repositioned on the surface. The values displayed are the maximum pressure recorded for each condition. The value is the maximum pressure reading of the average of 3 to 5 measurements of a 15 by 15 array of pressures for each condition.

DISCUSSION

Figure 1 demonstrates that the iterative optimization process achieved a support surface shape satisfying the established criteria. Initial and early iterations show that high pressure

[INSERT] Table 1. CASS Data from Pilot Experiment

[INSERT] Table 2. FSA Data from Pilot Experiment

measurements coincide with high stiffness measurements. After optimization, higher loads are shifted to support elements where low stiffness measurements are observed. As a measure of the degree to which this condition is achieved, the quantity of pressure times stiffness is observed. The lower the magnitude of this value, the closer the condition is to the optimal condition.

In Table 1, there is an apparent trend in the data with the lowest maximum pressure and stiffness values occurring for the optimal shape, and the largest maximum deflections also occurring for this shape. The maximum pressure (86.17 kPa) and stiffness (34.48 kPa/mm) measured on the CASS for the flat condition are values at the upper limit of the range measured by the pressure transducers and indicate that the actual values exceeded these quantities. Also in Table 1, the data indicating the pressure measured at the location on the seating array where stiffness was maximum illustrates that the algorithm performed as programmed and the inverse relationship between pressure and stiffness for the optimal shape was established. For the flat and initial shapes, maximum pressure is often, if not always, at the location where maximum stiffness is measured. For the optimal shape, the pressure is relatively low at these locations of higher stiffness, verifying that the optimization criteria has been satisfied. Again in pressure mat data in Table 2, a trend in the data is observed where peak pressures decrease from the flat foam to the initial contour and is usually lowest for the optimal contour.

CONCLUSION

The pilot experiments showed that the system's optimization procedure achieved a support surface shape satisfying the established criteria. The results also indicated that the optimal support shapes carved into high resiliency foam cushions provide lower peak interface pressures than flat foam and the support shape derived from the mechanical shape measurement system. The contours resulting from the mechanical shape system are based on the seated contour only and are not modified. The CASS system provides a useful tool for the investigation and quantification of the complex relationship among properties such as surface shape, interface pressure and tissue deformation under different load and stiffness conditions. This ability is of great significance for clinical study and application. Although the CASS itself is unlikely to be developed into a viable clinical tool for support surface design, the basic information gathered though its use may prove fundamental to successful design of custom contoured support surfaces. The information gained from the study of subject groups at high risk for pressure ulcers, such as the elderly, and persons with spinal cord injury or spina bifida, can be used to further refine and develop the control algorithm. The potential end result would allow application of the algorithm developed for a specific population to data taken from a passive mechanical shape measurement system in the clinic to create an optimized, custom-contoured support surface.

REFERENCES

[1] D.M. Brienza, K.C. Chung, C.E. Brubaker and R.J. Kwiatkowski. Design of a Computer-controlled Seating Surface for Research Applications. IEEE Transactions on Rehabilitation Engineering Vol. 1, No. 1, March 1993.

[2] D.M. Brienza, K.C. Chung and C.E. Brubaker, "Computer Design and Fabrication of Custom Contoured Seating," Medical Design and Material, Vol. 1, No. 1, pp. 32-41, January 1991.

ACKNOWLEDGEMENTS

This work was supported by a grant from National Institutes of Health, National Institute of Child Health and Human Development, National Center for Medical Rehabilitation Research, grant number R01-HD30161. Tricia Karg, MSBME Rehabilation Technology Program University of Pittsburgh UPARC-915 William Pitt Way Pittsburgh, PA. 15238 412-826-3138 tkarg@pitt.edu Evaluation of a Surface Shape Optimization Technique for Custom Contoured Cushion Design