Trajectory tracking for SCARA robots with compliant transmissions: A technique to improve the positioning precision

Mohd. Jawad Shaikh Fahed M. Sayed Nadeem Khan MHSSCoE, Byculla, Mumbai-8

Abstract: The paper describes thr robota project.The Robota project constructs a series of multiple degrees of freedom doll-shaped humanoid robots, whose physical features resemble those of a human baby. The Robota robots have been applied as assistive technologies in behavioral studies with low-functioning children with autism. Keywords: Trajectory tracking, compliance, SCARA robot.

II. Robota: Technological Developments A. Design considerations

The use of Robota as part of experiments with children with autism sets a number of constraints on the design of the robots mechanics and electronics. These are: The robot's size and weight must be sufficiently small i.e. the average size and weight of commercial dolls, namely 50cm in height and 1kg in weight. The processing must be done on-board and the robot should be battery-operated. The cost of the robot must be sufficiently low for the collaborating schools and museums to eventually be able to purchase it.

I. Introduction
Robota is the name of a series of doll-shaped minihumanoid robots. The Robota project is part of acurrent trend of robotics research that develops educational robot toys; see, e.g. [Michaud & Caron2002, Plaisant et al 2000, Kozima et al 2002]. Such a trend aims at understanding the role that anentertaining robot might play in the educational development of the child. This approach goesbeyond building playful robots, whose sole purpose is that of entertaining the user, e.g. [Fujita et al2003, Sawada et al 2004], by stressing the educational and therapeutic role of the robot [Dautenhahn1999]. The use of robots as entertainment systems puts important constraints on the design of both thephysical appearance of the robot and of its control system. For instance, the physical appearance of anentertaining robot should match a number of aesthetic characteristics to make it look playful and engaging [Breazeal 2002, Kitano & Fujita 2000; Montemayor et al, 2000]. In contrast, designing a robot for use with children with disabilities may, in some cases, put different constraints on the robots physical appearance. First, the robot will have to be robust beyond classical standards [Michaud & Theberge-Turmel, 2002] and it may not necessarily be advantageous to display human features. For instance, [Robins et al, 2004] showed that children with autism would interact more when the robots face is plain (no facial features) than when it bears a strong resemblance to that of a shuman.

III. Specification
Width:14 cm Height:45 cm Weight: 1500 g Degree of freedom: Head rotation, raising and lowering the left/right arm or left/right leg, coordinated motion of the two eyes (sideway) and individual motion of each eye (blinking) Actuators: DC Motors (Maxon A-max 26 mm) with clutches and position potentiometers

If we have to calculate the positioning error on a given trajectory, we need to consider the joint compliance of the manipulator; for this aim it is useful to define an elastic model of the transmission, for each degree of freedom (DOF) of the machine.

Table I. Parameters of SCARA robot

Fig.4 Model of an elastic transmission

Knowing the kinematic structure of the robot and its geometrical and physical parameters (lengths, masses, moments of inertia, stiffness and damping constants), the motion equations of the system can be deduced. The solution of the motion equations gives the positions velocity and accelerations of the robot links, when the motion of the actuators is assigned. The trajectory tracked by the end-effectors can be finally calculated by solving the forward kinematics for the position variables.
Table II. Shape parameters and mean positioning error before and after the optimization process

IV. Reduction of the trajectory tracking errors

Many industrial applications (cutting, welding, etc.) require a precise movement of the end-effectors along a given trajectory: in these cases the motion time is usually assigned, but the function s(t) (that defines the curvilinear abscissa of the end-effectors on the trajectory at the time instant t) is not specified: so it can be defined by the motion designer in order to optimize the machine performances. In this paper we will investigate if any particular function s(t) exists, that is able to reduce the trajectory tracking errors of the robot; the theoretical study can be carried out by using the dynamic model of the SCARA manipulator described in the previous paragraph.

In order to consider the joints compliance, the manipulator has been schematized by the elastodynamic model represented in Fig. 5 (see [6] for further details). The parameters used for the numerical simulations (see Table I) refers to a small SCARA robot, employed for pick and place operations; the geometrical and inertial data have been calculated according to the technical drawings of the machine, whereas the stiffness and damping constants of the joints have been experimentally identified, through free vibration tests carried out on an actual robot.

Fig 5 Schematic representation of SCARA robot

V. Conclusions
A method for improving the positioning precision of SCARA robots with elastic transmission have been presented in the paper. The proposed approach requires: A mathematical model of the manipulator that considers the joints elasticity. The use of a parametric law s = s(t), that defines the displacement of the end-effectors along the trajectory. The introduction of a performance index that gives a measure of the trajectory tracking error. An optimization procedure, which is able to determine the minimum of a multi-variable function. The minimization of the mean positioning error along the trajectory is achieved through a series of numerical simulations: a different tracking law s(t) is used for each trial and the best profile is selected by a suitable optimization algorithm.

VI. References
[1] Faglia R., Legnani G. Harmonic Drive transmissions: the effects oftheir elasticity, clearance and irregularity on the dynamic behavior of an actual SCARA robot. Robotica, vol.10, pp.369-375 (1992). [2] Faglia R., Legnani G. On the dynamic behaviour of robots with nonrigid transmission. 8th IFToMM Congress, Prague, Czechoslovakia (1991). [3] Cavalieri S., Martini M., Petrone F., Sinatra R. A neural network approach for position-error optimization problem in a redundant robot. 9th IFToMM Congress, Milan, Italy (1995). [4] Kim W., Rastegar J., Khorrami F. On the robot manipulator trajectory synthesis for minimal vibrational excitation. 5th IASTED Int. Conf., Cancun, Mexico (1997). [5] Rao S.S. Engineering Optimization - Theory and Practice. John Wiley & Sons, New York (1996). [6] Legnani G. Robotica industriale. Casa editrice Ambrosiana, Milano (2003) - in Italian.