A system used to establish the points of an industrial robot trajectory.

Stoica, Mihai ; Calangiu, Gabriela Andreea ; Sisak, Francisc 等

1. INTRODUCTION

Robot programming is a hot topic in the research field of robotics. Any robotic system must face this central difficulty: how to use an incomplete model in its environment to perceive, infer, decide and act efficiently? Oliver Lebeltel, Pierre Bessiere, Julien Diard and Emmanuel Mazer proposed a robot programming method that specifically addresses this question. They proposed a method to program robots based on Bayesian inference and learning. It is called BRP for Bayesian Robot Programming (Lebeltel et al., 2004).

The disassembly of used goods is characterized by strongly varying quantities and a wide range of different kinds and states of products. Furthermore, the requirements of a disassembly system are determined by the disassembly object, by the process as well as by the disassembly system itself. New methods for the programming of industrial robots have to be developed because the costs of programming small lot sizes are a key factor of the economic efficiency of small and medium sized enterprises (Uhlmann et al., 2008).

Optimal trajectory planning for robot manipulators plays an important role in implementing high productivity robots. The performance indexes used in optimal trajectory planning are classified into two main categories: optimum travelling time and optimum mechanical energy of the actuators. Most trajectory planning algorithms are designed based on one of the two performance indexes above. There are a few planning algorithms designed to satisfy two performance indexes simultaneously. On the other hand, there are some in the existing integrated optimisation algorithms of trajectory planning. In order to overcome those deficiencies, the integrated optimisation algorithms of trajectory planning presented are based on the complete analysis for trajectory planning for robot manipulators (Luo et al., 2004).

S. Mitsi, K.D. Bouzakis, G. Mansour, D. Sagris and G. Maliaris developed an off-line programming system which includes graphical simulation of the robot in its work cell, kinematic model of the robot, motion planning and creation of the NC code for manufacturing process. The proposed system is applied in a robot with five rotating joints for manufacturing operations and in a robot with six rotating joints for welding operations (Mitsi et al., 2004).

Programming of service robots is an expensive and difficult task especially when manipulator arms are involved. This is one of the drawbacks for everyday use of these systems. Programming by demonstration is a means to let users to program robots simply by demonstrating a task like putting something on a table or composing an object to a system that observes, interprets and then maps the performed user action to a given manipulator. Although progress has been made in certain areas, applications to real-world environments are limited (Zollner et al., 2004).

[FIGURE 1 OMITTED]

In this paper we want to present a system used to capture the points from the trajectory of a small industrial robot. The important piece of this system is a sensorial handle. In the second chapter we will show the design of the sensorial handle and in the third chapter we will present the system architecture.

2. DESIGN OF THE SENSORIAL HANDLE

The proposed sensorial handle is useful for small industrial robot programming, such as the Mitsubishi RV-2AJ. The design of this handle is presented in Fig. 1 and Fig. 2.

The sensorial handle consist of two cylindrical tubes (3 and 6 from Fig. 1; 1 and 2 from Fig. 2) placed in a concentric manner. The internal tube is connected by a magnet (2 from Fig.1) at the end-effecter (3 from Fig.1) of a robot. Between these tubes there are six sensors (six micro-switches) placed like as Fig.1 and Fig. 2. Around the internal tube; between the internal and external tube an elastic material is placed (7 from Fig. 1 and 3 from Fig. 2).

Practically, this handle is usefull for moving an industrial robot arm. After the robot arm is moved, the position can be memorized into a position table. For moving the arm, the human operator must push or pull the external tube into the direction in which he wants to move it. In this case the sensor from this direction is activated. For up and down movement sensors 5 and 4 from Fig. 1 are used. For ahead and behind movement are used sensors 9 and 8 from Fig. 1 and for right and left movement are used sensors 5 and 4 from Fig. 2.

[FIGURE 2 OMITTED]

Signals from these sensors are inputs of an electronics board, where they are computed. This board send commands to a PC. After that, the PC sends commands to the robot to move on the direction decided by the sensors. Because of the elastic material, if the external tube is free no sensor is activated. In this case the robot does not move.

3. SYSTEM ARCHITECTURE

The system architecture is presented in Fig. 3. It consists of four important elements: a sensorial handle (1), a robot (2), a PC (3) and an electronic board (4). The sensorial handle is mounted on the end-effecter of the robot. The signals from the handle are inputs for the electronics board.

The microcontroller (5), placed into the electronics board, computes these signals. It is connected with a PC via a serial link. On the PC runs a LabWindows CVI application. The application opens the serial connection with the microcontroller and sends commands to microcontroller to return the values from the handle's sensors. So, at this moment the microcontroller sends to the PC these values. The LabWindows CVI application also opens a serial connection with the robot too. Over this connection it sends commands to the robot. The robot moves according to these commands.

With this system we can capture the points from the end-effecter of a robot arm like the Mitsubishi RV-2AJ. Practically, we can establish only the OX, OY and OZ coordinates of the end-effecter. The robot arm works in two quadrants: for the OY coordinate the value is greater or equal than zero and for the OY coordinate of the end-efecter it is smaller than zero. When the robot arm passes from a quadrant to another, the software application sends a command to the robot controller to rotate with 180 degrees or -180 degrees the end-effecter.

[FIGURE 3 OMITTED]

The application has a graphical interface where the operator can fix the displacement precision and displacement value of the robot movement. On this interface a button is placed with which the operator can memorize, into a text file, the position of the robot. In this mode we can obtain the positions table. After that, the positions table must to be downloaded into the robot controller.

4. CONCLUSIONS AND FUTURE WORK

The system presented in this paper is useful for moving an industrial robot arm and memorizing points from its trajectory. It is very simple and can be used by an unskilled worker. With this sensorial handle the human operator can obtain the points from robot trajectory in a very simple manner. For that he must to move the handle and the robot moves in the same direction. The system has good results because the human operator can feel the movement of the robot. He must not know on which direction he must move the robot arm: whether the direction is on the OX, OY or OZ coordinates. With this system he only must move the handle in the desired direction. With this system the time for obtaining the robot points is small because the human operator can move the robot arm very quickly. We have tested the system on our robot: Mitsubishi RV-2AJ, but it can be used on other robots of the same type. The operator can use this system only if he has space around the robot to move the handle. Also it can be used only for small robots. If the robot is big, the human operator does not seize the handle in some situations.

The system has good results but with this sensorial handle (presented in chapter 2) only the OX, OY and OZ coordinates in the space can be obtained. With this handle the human operator can not rotate the end-effecter in XOY plane and can not rotate it contorted by OZ axis. In our future work we want to improve the sensorial handle in order to realize these rotations. We want, in our future work, to remove the PC from system architecture. We want to connect the electronics board directly to the robot controller.

5. REFERENCES

Lebeltel, O.; Bessiere, P.; Diard, J. & Mazer, E. (2004). Bayesian Robot Programming, Autonomous Robots, Vol.16, No. 1/January 2004, pp 49-79, ISSN 1573-7527

Luo, X.; Fan, X.; Zhang, H. & Chen, T. (2004). Novel Integrated optimization algorithm for trajectory planning of robot manipulators based on integrated evolutionary programming, Journal of Control Theory and Application, Vol. 2, No. 4/November 2004, pp 319-331, ISSN 1993-0623

Mitsi, S.; Bouzakis, K.D.; Mansour, G.; Sagris, D. & Maliaris, G. (2004). Off-line programming of an industrial robot for manufacturing, The International Journal of Advanced Manufacturing Technology, Vol. 26, No. 3/August 2005, pp 262-267, ISSN 1433-3015

Uhlmann, E.; Friedrich, T. & Bayat, N. (2008). Development of a technology orientad programming system for industrial robots, Production Engineering, Vol. 2, No. 1/April 2008, pp 103-108, ISSN: 1863-7353

Zollner, R.; Rogalla, O.; Ehrenmann, M. & Dillmann, R. (2004). Mapping Complex Tasks to Robots: Programming by Demonstration in Real-World Environments, In: Advances in Human-Robot Interaction, Springer Berlin, ISBN: 978-3-540-23211-7, Berlin