Control and design for biomimetics application using smart materials.

Bizdoaca, Nicu ; Tarnita, Daniela ; Petrisor, Anca 等

1. INTRODUCTION

Biomimetics is a new multidisciplinary domain that includes not only the uses of animal-like robots--biomimetic robots as tools for biologists studying animal behaviour and as research frame for the study and evaluation of biological algorithms and applications of these algorithms in civil engineering, robotics, aeronautics. Life's evolution for over 3 billion years resolved many of nature's challenges leading to solutions with optimal performances versus minimal resources. This is the reason that nature's inventions have inspired researchers in developing effective algorithms, methods, materials, processes, structures, tools, mechanisms, and systems.

A promising field in practical implementation of biomimetics devices and robots is the domain of intelligent materials. Unlike classic materials, intelligent materials have physical properties that can be altered not only by the charging factors of that try, but also by different mechanisms that involve supplementary parameters like light radiation, temperature, magnetic or electric field, etc. These parameters do not have a random nature, being included in primary math models that describe the original material. The main materials that enter this category are iron magnetic gels and intelligent fluids (magneto or electro-rheological or iron fluids), materials with memory shape (Humbeeck, 2001) (titan alloys, especially with nickel), magneto-electric materials and electro-active polymers.

2. BIOMIMETIC EXPERIMENTAL PLATFORMS

At the Department of Mechatronics, Faculty of Control, Computers and Electronics, University of Craiova, graduates and PhD students are implied in projects regarding biomimetic robotic structures.

The key aspects of projects concern the movements' similarity with movements of biological counterparts, obtained using particular mechanical and kinematical architecture and especially control programs implemented in individual control architecture. Special attention was given to tentacle (trunk) robotic structure (Ivanescu, 1986).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

3. CONVENTIONAL CONTROL PERFORMANCES

In order to investigate performances of conventional control on SMA robotic tentacle (Bizdoaca & Diaconu, 2006) unit comportment, a Quanser modified platform was used for experiments. The basic control structure uses a configurable PID controller and a Quanser Power Module Unit for energizing the SMA actuators. PID controller was changed, in order to adapt to the particularities of the SMA actuator. A negative command for SMA actuator corresponds to a cooling source. The actual structure do not use for cooling other devices, excepting the ambient temperature.

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

Using PID, PD controller the experiments conduct to less convenient results from the point of view of time response or controller dynamics.

The best results arise when a PI controller is used.

The PI experimented controller parameters are: the proportional parameter [K.sub.R] = 10 and the integration parameter is [K.sub.I] = 0, 05. The input step is equivalent with [30.sup.0] angle base variation and the evolution of this reference is represented by the response of the real system in Figure 14. The control signal variation is presented in Figure 15. Using heat in order to activate SMA wire, a human operator will increase or decrease the amount of heat in order to assure a desired position to robotic module. Because of medium temperature influence, it cannot be established, apriori, a clear control law, available for all the points of the robotic structure workspace.

4. DESIGN OF THE FUZZY LOGIC CONTROLLER

The high level hierarchical control problem asks for determining the torques T so that the trajectory of the overall system (object and manipulator) will correspond as closely as possible to the desired behaviour (Ivanescu& Bizdoaca, 2000). The controller receives the error and the change of the error components, [e.sup.i], [[??].sup.i] for each units of the tentacle manipulator and depending on the values of forces [[tau].sub.Fi], generates the fuzzy control torques [T.sup.i.sub.F]. The control system contains two parts: the first component is a conventional controller which implements a classic strategy of the motion control based on the Lyapunov stability and the second is a Fuzzy Controller.

The control rules are determined by the motion in the neighbourhood of the switching line as a variable structure controller. We adopted here a special class of SMC (Fig 14) named DSMC (Fig.15) (Direct Sliding Mode Control) (Ivanescu& Bizdoaca, 2003).

[FIGURE 13 OMITTED]

[FIGURE 14 OMITTED]

[FIGURE 15 OMITTED]

[FIGURE 16 OMITTED]

The physical meaning of the rules is as follows: the output is zero near the switching line, the output is negative above the switching line, the output is positive below the diagonal line, and the magnitude of the output tends to increase in accordance with magnitude of the distance between the switching line and the state. The state space of e, [??] will be partitioned initial into nine fuzzy regions (negative--N, zero--Z, positive--P, with trapezoid membership function). The fuzzy if-then rules for these fuzzy regions are presented in Table 1.

The control assures the motion of the system on the first part of the trajectory with ki SMALL. When the trajectory penetrates the switching line, the DSMC is applied by the control of the coefficient ki, with ki is BIG. If the evolution is not satisfactory, a new control strategy is adopted. The finer fuzzy domains are introduced and new fuzzy partitions are used: big negative (BN), small negative (SN), negative zero (NZ), zero (Z), and positive zero (PZ), small positive (SP), big positive (BP).

Fig. 15 represents the trajectory in the plane [??], e for fuzzy SMC procedure and Fig. 16 the same trajectory for a DSMC procedure for a new switching line. We can remark the error during the 1th cycle and the convergence to the desired trajectory during the 2nd cycle.

5. CONCLUSION

In this paper a Direct Sliding Mode Fuzzy Controller for SMA tentacle robotic structure is applied. The controller is tested using a model of the tentacle structure. Further research will be focused on experimental tests using fuzzy control to our experimental SMA tentacle structure. The fuzzy controller offer better performances in term of the repeatability of the results (precision, time response) without ambient temperature control.

6. ACKNOWLEDGMENT

This research activity was supported by Ministry of Education, Research and Innovations, PNCDI 2--289/2008 Reverse Engineering in Cognitive Modelling and Control of Biomimetics Structure

7. REFERENCES

Bizdoaca N.G., Diaconu I.(2006) Hyperredundant Shape Memory Alloy Tendons Actuated Robotic Robot, ICCC 2006, pp. 53-56, ISBN 80-248-1066-2, Czech Republic

Humbeeck J. Van (2001), Shape Memory Alloys: A Material and a Technology,pp.837-850, Adv. Eng. Mater, v3 i11

Ivanescu M. (1986), "A New Manipulator Arm: A Tentacle Model", Recent Trends in Robotics, pp. 51-57, ISBN:0444-01140-4

Ivanescu M., Bizdoaca N.G.(2003), Dynamic control for a tentacle manipulator with SMA actuators, IEEE International Conference on Robotics and Automation, pp. 2079-2084 ,ISBN 0-7803-7737-0, Taipei, Taiwan

Ivanescu M., Bizdoaca N.G., An Intelligent Control System for Hyperredundant Cooperative Robots (2000), Proceedings ISRA 2000, pp. 214-220, Monterrey, Mexico

Tab. 1. The initial
fuzzy if-then rules

[??]\e N Z P

P Z N N
Z P Z N
N P P Z

Tab. 2. The finer fuzzy if-then rules

[??]\ e BN SN NZ Z PZ SP BP

BP Z NZ SN SN BN BN BN
SP PZ Z SN SN SN BN BN
PZ SP SP Z NZ SN SN BN
Z SP SP PZ Z SN SN SN
NZ BP SP SP PZ Z NZ SN
SN BP BP SP SP PZ Z NZ
BN BP BP BP SP SP PZ Z