Mechatronic design algorithm for human prostheses: intelligent robotic end effectors.

Dolga, Lia ; Dolga, Valer

1. INTRODUCTION

Robotic- and medical field require both advanced physical gripping systems. As they are multi-technological products, the mechatronic design philosophy of approaching complex systems is suitable to develop or improve design solutions. Moreover, biomechatronics integrates mechanical elements in the human body, allowing describing, analyzing, designing and improving any group of objects that work jointly to generate an expected result destined for a live system. This could be a single organism, organ, or any sophisticated combination of artificial and alive components.

The authors of this paper studied the similarities between the human hand behaviour, the prosthetic- and the industrial gripper, highlighting the common aspects.

2. A HUMAN HAND BIOMECHANICAL STUDY

A complex configuration, with special parameters, characterizes the human hand biomechanical "actuator", since the mechanism that transmits the force includes joints with an uninterrupted friction. Friction forces are low; the inertia of the parts is also low. Dead stroke and lost motion are absent.

The biomechanical analysis of the muscles envisages the forces, the elasticity and the hysteresis and is essential for creating artificial systems very similar to the human hand.

The muscle fibres are single cells of variable length. A muscle's length-tension curve illustrates how its tension comes from two sources (Figure 1) (Thompson, 2001): active- and passive tension. Active tension derives from the interaction between actin and myosin fibres and has a nonlinear variation with the length of the muscle. Passive tension can develop in the muscle's complex connective tissue; this dependence is either linear (Pons 2008) or parabolic (Thompson, 2001).

[FIGURE 1 OMITTED]

3. CHECK-LIST FOR THE DESIGN ALGORITHM

The grippers design specifications are either individual collections or a set of collections and represent the checklist. The authors propose checklists that correspond to the organic level from the mechatronic systems theory. A specific checklist rules each design stage. Regardless the accepted type of actuator, one has to answer:

* Which generalized force gives the required gripping force?

* Which are the geometric parameters of the motion, the imposed speed and the additional speed restrictions?

* Which is the allowed energy upper limit? ...

* ... Which is the actuator breakdown effect on the prosthesis? How is redundancy guaranteed, even in dynamic cases?

* Did the designer explore any actuator type: piezoelectric, electromagnetic, shape memory alloy, magnetostrictive, thermal expansion, hydraulic, pneumatic, etc?

The inserted mechanism plays a double role: to allow increasing the distance actuator- driven zone and to convert the output motion of the actuator in a rotation motion of the fingers. The designer has to solve further problems (the number of the independent motions, the limited working space, the presence of passive degrees of freedom in the fingers structure, etc).

The sensorial system (proprioceptive- and exteroceptive subsystems), is vital for an intelligent prosthesis or robot.

Designing the control system requires to select either a multilevel control system or an intelligent system, to detect software errors and prevent power failure effects.

The designer has to protect the prosthesis, the cables and the connexions against electromagnetic interference.

4. EVALUATION, OPTIMIZATION SELECTION

Let [V.sub.i] (i = 1, 2 ... m) be the set of the achievable design variants as a result of the dimensional design. Multi-objective decision-making optimization methods reveal a new collection of variants, dimensionally optimized. The variants in the last collection are compared with respect to a set of criteria. The multi-attribute decision-making process selects the best variant. The decision making process evolves in three stages.

The developed models admit various representations: abstract or concrete, in brief, or in depth. The systemic approach of the task is essential. Several methods of new ideas development are helpful, like "the morphological chart" (Van der Hoog et al., 2009). The method starts from the secondary functions developed around the main functions of the systems.

4.1 Case study 1

For a prosthesis with a c. c. servosystem and a c. c. variator, Table 1 presents the secondary functions associated to the variator. Table 2 displays a fragment of the morphological chart. From the obtained matrix (functioni, [sub-solution.sub.1,j]), the set of satisfactory sub-solutions for future analyses is selected.

The integration of actuators and dedicated sensors for a particular purpose is a basic practice in mechatronic systems. A functional principle can lead to variants of dedicated systems.

Table 3 contains several basic configurations of actuators with shape memory alloy. Typically, the driving kinematic pairs are either a rotation kinematic pair, or a translation kinematic pair. Table 3 illustrates two variants of actuators: bias- spring and differential. The former employs one piece of wire, from shape memory alloy, coupled with a bias spring; the latter includes two pieces of wires in differential mode.

The second stage, the constructive and functional optimization of an organ variant, primarily applies the classical procedures within the mechanical- or the electrical field, depending upon the type of the analyzed component.

The third stage considers a finite set of comparison criteria, appropriate to the evaluated organ, and applies multi-attribute decision methods to compare a finite set of optimized variants, determined during the second stage (Resteanu et al., 2006). The criteria can be quantitative (values) or qualitative (levels) and each criterion may try to attend a maximum or a minimum.

4.2 Case study 2

One aims to select the best type of actuator for a mechatronic system that performs investigations. Three variants of actuator are available, with ten performance criteria, quantitative or qualitative, pointing to a maximum or to a minimum (Efficiency, Power to weight ratio, Force to cross-section area, Durability, Stiffness, Overload ratio, ...).

The evaluation of each variant [V.sub.j] (j=1, 2, 3) through each criterion Ci (i=l ... 10), is reflected in the consequence matrix A, which was inhomogeneous and was homogenized by a vectorial normalization. The normalized matrix is R = ([r.sub.ij]).

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

[r.sub.ij] = [a.subij]/[square root of [[summation].sup.n.sub.1][a.sup.2.sub.ij]] (2)

Qualitative criteria were transformed using levels of conformity, to become quantitative.

[TABLE 2 OMITTED]

[TABLE 3 OMITTED]

Since criteria are more or less important, a scale of relative weight is suitable. The study uses 10 criteria; the vector of the coefficients of importance, P, and the importance matrix B are:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

The TOPSIS multi-attribute decision method (Resteanu et al., 2006) is applied to find the proximity of variants to the ideal solution (Table 4). A hierarchy of the variants, in a descending order of [K.sub.i] ([V.sub.1], [V.sub.2], [V.sub.3]) is established. The best variant is [V.sub.1].

5. CONCLUSION

The adoption of a mechatronics design approach brings benefit. This is both in terms of added functionality for the same design costs and a reduced price for similar functionality when compared with a gripper designed by a conventional approach. Regardless the level of technology, simple or advanced, the motivation in adopting a mechatronics approach for grippers design provides reduced development time and costs, efficiency and functionality for manufacturing.

The authors plan to extend future research over new cybernetic solutions for grasping devices within arm prostheses, where the subject becomes even more complex.

6. REFERENCES

Pons, J. L. (2008). Wearable robots: biomechatronic exoskeletons, John Wiley & Sons, Ltd, West Sussex, ISBN: 978-0-470-51294-4

Resteanu, C; Somodi, M & Alexe, B. (2006). Multi-Attribute Decision Making E-Course, Proceedings of the International Multi-Conference on Computing in the Global Information Technology (ICCGI'06), pp 11-11, ISBN: 0-7695-2690-X, Bucharest

Thompson, D.M. (2001). Muscle anatomy and function, http://moon.ouhsc.edu/dthompso/namics/know.htm, Accessed on: 2009-05-26

Van der Hoog, W.; Van Boeijen, A.; Van de Geer, S. & Tassoul, M. (2009). Morphological chart, The industrial design engineering wiki, http://www.wikid.eu/ index.php/Morphological_chart, Accessed on 2009-05-10

***(2004). Prehension overview. http://www.bsu.edu/web/ jkshim/handfinger/overview/, Accessed on: 2009-04-22

Tab. 1. Secondary functions of a c. c. variator

 FUNCTIONS

Power Voltage Speed Control
control regulator ... measuring architecture

Tab. 4. Results of the TOPSIS method in the studied case

Variant Proximity of the variant to the ideal solution

V1 [K.sub.1] = 0,58
V2 [K.sub.1] = 0,39
V3 [K.sub.1] = 0,42