VI Congreso Internacional de Ingeniera Mecnica y IV de Ingeniera Mecatrnica IV Congreso Internacional de Materiales, Energa y Medio Ambiente Bogot, Colombia.

2, 3 y 4 de mayo del 2013

Smooth velocity profile analysis for a three degree of freedom parallel robot Delta type, developing linear paths on a virtual Delta kinematic Simulator on Autodesk Inventor.
Santiago Rodrguez*, Ubaldo Garca, Jorge Garca , Universidad Santo Toms, [email protected] Universidad Santo Toms, [email protected] Universidad Santo Toms. [email protected]
*

ABSTRACT
It's common to find nowadays that the jerk analysis is so important on any actuator and specially on machines that has high precision on its moves, the velocity profile has couple of years of studies and has evolved to a point that its indispensable to develop trajectories in any working volume [1]. For years, some important companies involved to the high precision industrial machines like ABB and Adept, has develop lots of parallel type machines like the robot delta and they have reach the level of high precision and high performance with good control of speed and acceleration. A delta robot type machine is a precision robot that makes easy every kind of placement in just seconds, this is possible because of its architecture, but it has to be handle by the control of the velocity profiles, which permit the management of the velocity ramp, making them smooth. The phenomenon of jerk also tends to a finite value; this means that it is riskless for the actuators in the way that the jerk causes failures when the smooth ramp of velocity goes from zero to the velocity value [2].

Key Words: End effector (Efector Final), Delta Robot (Robot tipo delta), Jerk (Jerk), Speed (Velocidad), Acceleration (Aceleracin).

INTRODUCTION

This article aims to show the effects of velocity, acceleration and torque when a delta robot is acting in a pick and place application, analyzing the velocity profiles mathematics on a Kinematic Simulator made on Autodesk Inventor. The velocity profile is a velocity ramp, in this Case is smooth and permits to raise the velocity from zero to a certain speed on a period of time. Controlling the velocity its possible and that is going to affect the behavior of the acceleration, and the jerk[4]. The control of velocity is really interesting and is one of several factors that affect the motion on every high precision machine, including robots of every single architecture, and CNC manufacturing machines. Controlling the velocity of work at the end of the effector of the Delta, and knowing the masses and the accelerations of the system its possible to calculate the maximum torque that the manipulator will suffer when it reach the maximum peak velocity. Torque calculation its indispensable because it helps to select the proper actuators, with or without gear reduction, and also a couple of characteristics that are essential like: Current, Voltage, max torque, max rpms. The virtual manipulator is the one how is going to prove how its possible to determine the torque that its present when the end of the 3 DOF delta is moving or doing any path on the working volume. The digital simulator will predict as accurate as it could fit mathematically the answer of the torque doing direct Kinematics and this will be compare with the inverse Kinematics done in Matlab software[5].

MATERIALS & METHODS

2.1

Smooth Velocity Maths

The Smooth velocity profile merges the trajectory segmented points but with a different sense of action, because now those points on every axis coordinate moves with a different velocity in the same lapse of time[1]. The importance of the Smooth velocity profile is that rather than moving the end effector point by point with the same velocity, we have a time of Smooth acceleration, then a gap of constant velocity and finally a Smooth desacceleration movement, all in the same lapse of time. The speed up process its described in the equation (1) all assuming that it moves in the 3 axis at the same time in a determined lapse of time that is calculated with a speed given by the user.

(1)

For this case the main purpose is to get position equations that may generate linked paths with Smooth velocity parameters, from that, we integrate the velocity equations and we get these equations as list in equation (2).

(2)

; The constants of the equations are the form equation slope that might have every position on the different axis. All the constants are calculated like in equations (3). The equations (4) describe the position equations on the graph.

(3) ; ; ;

(4)

The constant velocity its reached in the second segment of the graph, this is described in the equation (2) with .

2.2

Smooth velocity profiles Analysis

The analysis consist on simulate a normal path on the working voume, mounting all the data to the simulator of the Kinematics 3DOF delta manipulator. This articule will expose step by step the process of making the analysis of Smooth velocity profiles with mathlab .EXE and the virtual manipulator made on Autodesk Inventor [5].

Step 1. Define the velocity profile speed, for this case is going to be used

as shown on

the Smooth velocity profiles calculator. With a trajectory showed in figure (1), in which is define the initial spot of the trajectory and previously the end point.

Figure 1. Linear path of the trajectory implemented.

The position graph in figure (2) shows for every actuator how it looks like when the Smooth velocity profile calculates the position with in a speed of 16 .

Figure 2. Position graph for all axis using mathematical approach.

To get an idea of how the velocity profiles and the torque behaves when the robot is interpolating a linear path with the smooth velocity profile theory, the virtual simulation tool in Autodesk inventor gives a solution for the torque behavior and this makes easy to compare the if the profiles are accurate enough to decrease the impact of the acceleration and torque on the system.

Step 2. Exports data from Smooth profile calculator to the Kinematics solver by .TXT files in which the order of the data is important, because this is the order of reading the content of the data by the Inverse Kinematics solver and the Forward Kinematics simulator shown in figure (3) and (4). Step 3. Pass over the Smooth profiles data to the Kinematics solver and this will exports 3 files that are the main data files from every single actuator to pass it over the kinematics simulator. Figure 3.Inverse Kinematics Solver.

Figure 4.Inverse Kinematics position graphs.

Step 4. Simulate the Smooth profiles on the Kinematics Simulator and see the results for the Smooth profiles graphs like Speed and position.

3 3.1

RESULTS

Velocity Profiles Compared on autodesk inventor an matlab.

Position graphs shown with the mathematical equations of the Smooth velocity profile are exactly the same as the result that plots Autodesk inventor on the solution, this is showed in figures (5) for X axis, (6) for Y axis, (7) for Z axis and (8) the plot of the three graphs on inventor respectively.

Figure 5. Position graph for X axis, Simulation approach.

Figure 6. Position graph for Y axis, Simulation approach.

Figure 7. Position graph for Z axis, Simulation approach.

With the illustration of the graphs its easy to know that in a linear path, the delta will move two of the three motors to the same direction but with a different speed and a different magnitude, and the other one its turning around to the opposite side.

Figure 8.Position Graphs from the path with Smooth velocity profile.

DISCUSSION

The article describes the process and makes emphasis on the movement of the 3DOF manipulator and focus the important of the Inverse kinematics solver is able to do, as it is known, the inverse Kinematics solves for the inverse and in this case the kinematics simulator is correcting the calculation done, because this one uses direct kinematics. In other words, what is done on inverse kinematics, the simulator verifies, and with the result is important to say that the solver is accurate and that the velocity profiles calculated on matlab are exactly the same on inverse simulation and in direct kinematics simulation. The values of the position obtained with inventor and the position graphs obtained in matlab are the same excepting in Z axis because it seems that inventor uses a regression to fit the values, so the way to fix this is solving the position for more velocity values.

Its accurate to say that if the motion process is simulated well and all the results are true, the possibility of selecting the actuators using this tool is possible and taking into account that it makes a process in which is simulated all the variables that determines the motion of the 3DOF manipulator, so this is the most accurate tool nowadays to select an actuator by torque analysis using Smooth velocity profiles.

REFERENCES

[1] J.A. Garca and E. Crdoba,Generacin de trayectorias rectas, utilizando perfiles suavizados de velocidad, en mquina con arquitectura paralela pkmr -3dof [2] D. Constantinescu and E. a Croft, Smooth and time-optimal trajectory planning for industrial manipulators along specified paths, 1998. [3] F. Zou, D. Qu, and F. Xu, Asymmetric s-curve trajectory planning for robot point-to-point motion, 2009 IEEE International Conference on Robotics and Biomimetics (ROBIO), Dec. 2009, pp. 2172-2176 [4] J.A. Garca and E. Crdoba, Trajectory Generation Control In The PKMR Machine -Robot of Orthoglide Type, 23rd ISPE International Conference on CAD/CAM Robotics and Factories of the Future, 2007, p. 8. [5] U.Garca Torque Analysis of a Three Trablational Degrees of Freedom Parallel Manipulator 23/01/2012. [6] L.W. Tsai. Robot analysis: the mechanics of serial and parallel manipulators. A WileyInterscience publication. Wiley, 1999. ISBN 9780471325932.

[7] L.-W. Tsai, Robot Analysis, John Wiley & Sons, Inc., 1999

10

[8] W. Younis. Up and running with Autodesk Inventor Simulation 2010: a step-by-step guide to engineering design solutions. Butterworth-Heinemann/Elsevier, 2009. ISBN 9781856176941. [9] S. Macfarlane and E. a. Croft, Jerk-bounded manipulator trajectory planning: design for real-time applications, IEEE Transactions on Robotics and Automation, vol. 19, no. 1, pp. 42-52, Feb. 2003. [10] K. D. Nguyen, I.-ming Chen, and T.-chew Ng, Planning algorithms for s-curve trajectories, 2007 Ieee/Asme International Conference on Advanced Intelligent Mechatronics , vol. 1, pp. 1-6, 2007. [11] D. Constantinescu and E. a. Croft, Smooth and time-optimal trajectory planning for industrial manipulators along specified paths, The University of British Columbia, 1998. [12 K.-H. Rew, C.-W. Ha, and K.-S. Kim, A practically efficient method for motion control based on asymmetric velocity profile, International Journal of Machine Tools and Manufacture, vol. 49, no. 78, pp. 678-682, Jun. 2009.]