3/ Material Handling and machine assistance

Introduction
Component handling is an ideal application for an industrial robot. It is a repetitive operation, carried out often under unpleasant and hard working conditions, which requires little skill. Component handling tasks usually involve fairly simple manoeuvres, with only modest accuracy being required. It may also include associated ancillary operations such as applying lubricants or air blast cleaning. These operations can be perfectly synchronised within efficient cycle times and during high-volume, repetitive production runs.
A handling process consists of eight sequences:
1)  Transfer of the robot arm up the workpiece
2)  Fine motion approaching the workpiece
3)  Grasping
4)  Fine motion uprising the workpiece
5)  Transfer the workpiece to the desired position
6)  Fine motion down to the destination position
7)  Release the workpiece
8)  Fine motion upward
The gripping sequence is the more delicate part of handling material. The gripper's and the robot's positions have to be checked for collision with other objects in the environment. The limitation of the robot's workspace can also be a problem for all other planning sequences of the handling.
Gantry robots, Fig. 1, are often used for material handling, machine assistance, or palletising and depalletising; they offer advantages as large work envelopes, large payloads and flexibility, often one gantry robot handles many machines and variety of parts; the main constraint is that it requires high ceiling (over 3m).

Figure 1: Gantry robot for machine assistance
Main functions for a material handling modular robotic system:
- High volume, multi-position palletising
- Depalletising.
- Order-picking
- Display (mixed) pallet building
- Automated storage & retrieval
- Loading and unloading of machine tools
- Part transfer

3.1/Material Handling

Many component handling applications involve pick and place operations, Fig.2, packaging/unpackaging or palletising/depalletising of finished products, Fig. 3. Characteristics of palletising/depalletising and packaging/unpackaging operations often require:
a)  Moderate accuracy.
b)  Simple geometry and control.
c)  High load-carrying capacity.
d)  Good reach and mobility.
e)  Sympathy with delicate cargoes.
f)  Variety of gripping methods.
 
 
Figure 2: ABB material handling robot with vaccum gripper.		  
Figure 3: ABB depalletising robot
3.1.1/ The gripper problem

The grasp planning might fail because of an inadequate gripper. If so, the gripper must be changed for the actual handling task.
Handling non-rigid product requires special grippers. These products can be easily deformed during handling action and it is necessary to have the capacity to manipulate the object within the gripper, to control an imposed deformation or minimise deformation as appropriate. A fundamental aspect to consider is the use of sensing systems and close loop control of actuators [Abr92]

3.1.2/ The Release Problem [Rei95]

With this problem none of all possible grasp configurations of the actual gripper allows a collision-free release of the object. The handling might be done by dividing the handling process into two succeeding handling processes with an intermediate position Fig. 4, the object's initial orientation and the grasp configurations found for the final position have to be considered.

Figure 4: Two succeeding handling processes with an intermediate position

3.1.3/ Robot vision system [Rts97]

Workpiece vision recognition, Fig. 5, is the more common solution in industrial handling applications. The workpiece is viewed by a camera system and its orientation is determined by a recognition 3D system, all with flexibility, high speed and fast processing abilities.

Figure 5: Example of vision recognition in material handling task, the shape of the products
on the conveyor is identified and located by a camera. After acquisition by the vision system the robot picks the products and puts them into the predefined position. Broken products are rejected. A : Vision camera. B : Table top robot. C : Gripper. D : Camera field of view. E : Conveyor. F : Vibration feeder.

A vision system for robotised material handling tasks is able to capture a digital image of the picking scene and locate, and identify a workpiece according to its size, form, marks, colouring, and text strings. It can determine the location, angle and scale of the scene as necessary.
The material handling applications that need the more vision recognition systems are the unloading and loading of pallets and the picking of workpieces from pallets and conveyors. This technology offers many benefits including, greater accuracy, lighter mechanics and fast product changes.

3.1.3.1/ Workpiece recognition problems [Ble96]:

Applying vision-based object recognition methods in a manufacturing environment is affected by a wide range of disturbing factors, e.g. reflections and shadows due to specific illumination conditions and surface characteristics, or objects, from which only a fraction is visible, Fig 6, this may lead to a complete failure of the object recognition unit.

 
Figure 6: Typical problems encountered by vision-based systems in a manufacturing environment.
The usage of illumination can make the recognition easier. The most common and simple methods  are the front and the back lighting. To gain better contrast of the picture, back lighting should be used, because in this case the darker shape of the part can be separated easily from the lighter background. In the case of front lighting the good contrast is lost, but more information can be acquired from the part.

3.2 / Machine assistance [Shi85]

The typical operations of a machine assistance robot are loading and unloading of workpieces and machine tools.
A robot is normally designed to pick up a part in the same attitude and position each time; therefore proper part preparation for robot handling is required. Jigs and fixtures are commonly supplied to provide accurate part placement for this purpose. In less common situations, machine tools are loaded with parts by vision-equipped robots, which can locate parts that are placed at less accurate positions. A machine-mounted robot is designed to swivel its arm to load a part for one machine operation, then set the part down and repick it at a different orientation for the next operation. Fixtures are used to hold and maintain parts positions during these motions. Robots can be useful for most machine-loading operations, however when high precision processes are required, special care must be taken. Typical positioning accuracy of machine-mounted robots is ±0,05 mm. Robots may not be justified when parts require extremely close tolerances that are beyond the accuracy of the robots. Part unloading and removal from a machine is just as important as supplying raw material and loading new parts. Various alternatives exist for part removal, including removal by square wooden pallets, roller conveyors or removal by robot carts.
Machine-mounted robots have often a long cycle time. An extreme example shows that a machine-mounted robot requires about 18 sec. to load a chuck, compared to 5 sec. by a human loading (about 4 kg) workpiece.

3.2.1/  Press shop assistance [Mül92]

Industrial robots are well-suited to fulfilling the high requirements involved in interlinking, loading and unloading individual presses. The main requirements are:
a)  Short cycle times.
b)  High availability.
c)  Short retooling times.
d)  High flexibility (short start-up and programming times for production of new parts, and Fast interchangeability of
     components.
e)  Rugged construction of the automation equipment (because of vibration).

The principal difficulty to be overcome in handling the workpieces is posed by the high acceleration forces acting on the panels during transfer. Complex press tools with the position of  the workpiece in relation to the press centre and the angular orientation varying from press to press are common. Complicated loading motion are often necessary, with the workpiece having to be turned by up to 90 degrees against the direction of transfer. Handling the workpiece within the press can be further impeded by lateral dies on the lower press tool and by guide elements. Between two presses the workpiece often has to be turned over. The panels are often heavily oiled, which makes it difficult for them to be handled with suction mechanisms. In this cases, the advantages in terms of flexibility offered by 6- or 7- axis robot other automation variants take full effect.
Press-linking robots can be universally implemented for the following tasks:
a)  Full mechanisation
b)  Partial mechanisation
c)  Lateral removal from the line
d)  Unloading of the last press
e)  Combination of feeder unloading and interlinking with robots presses are spaced very far apart (>9m)

The use of  a 6-axis robot for interlinking individual presses offers the following advantages over a 7-axis robot:
- The higher payload capacity of the robot allows larger and heavier components to be handled.
- The capital cost is reduced owing to the omission of the seventh axis.
- On account of the kinematics of a 6-axis jointed-arm robot, the panels are guaranteed to be moved along a linear path during transfer between the presses, normally close to the centreline of the press line.

Remark: In this application the 7th axis allowed to the robot the displacement from one press to another (circular trajectory).

Another important problem that a press assistance robot may face is vibration. The allowed tolerance for the robot environment vibration is 1G; often robots are exposed to 3G till 7G in a press shop. In that case an anti-vibration device must be provided to prevent the vibration being transmitted to the robot body. A rubber is usually used for floating the robot on the press body [Oga93].

3.3/ A survey how the robot performance affect the quality

· Repeatability: it has to be high if the robot is not using a vision recognition system (e.g. the parts are always presented with the same orientation and in the same position).
· Absolute accuracy (off-line programming, vision guidance).
· The static compliance: it must be low (high stiffness) especially when the robot is manipulating different kinds of heavy loads as in foundry or press shops.
· Path accuracy: some tasks need a high path accuracy, as handling parts in press shops with particular trajectories to avoid obstacles.
· Cycle time: it must be short to operate with a high velocity.
· Number of axis: the 7th axis is not always suitable.

3.4/ A survey how other parameters affect the quality

· The gripper (unsuitable gripper, lack of flexibility).
· Vision system (error in recognition, bad positioning of the camera, bad calibration, bad illumination, lack of accuracy).
· Programming errors.
· External environment conditions (vibration in press shops, high temperature in foundries).
· Other peripheral equipment (conveyors speed, fixtures, feeders jamming, pallets).

3.5/  Proposal of a ranking list of performance criteria

1.  Pose repeatability - repetitive task.
2.  Pose accuracy - off-line programming, vision guidance.
3.  Path repeatability.
4.  Path accuracy - off-line programming.
 

References

[Rei95] G. Reinhart, D. Kugelmann, Robot Handling in Case of Disturbances with 3D Simulation, Institute for Machine
            Tools and Industrial Engineering (iwb) Prof. Dr.-Ing. G. Reinhart, Prof. Dr.-Ing. J. Milberg Munich University of
            Technology 85609 Aschheim, Germany, Proceedings of the Third IASTED International Conference on
            Robotics  and Manufacturing 1995, CancÇn, Mexiko.
            http://www.iwb.mw.tu-muenchen.de/~kugelman/klsfb331/cancun:Dg.html.

[Abr92] P. Abreu, P.N. Brett, Robotic Gripper for the Manipulation of Non-rigid Products, Advanced Manufacturing
            and Automation Research Center, University of Bristol, 23rd International Symposium on Industrial Robots,
            Barcelona 6th-9th October 1992.

[Ble96] S. Blessing, S. Lanser, C. Zierl,: Vision-based Handling with a Mobile Robot. In: M. Jamshidi, F. Pin, P.
            Dauchez, Proceedings of the Sixth International Symposium on Robotics and Manufacturing (ISRAM'96)
            Montpellier, France 1996, Recent Trends in Research and Applications Volume 6, ASME Pres, New York,
            1996.

[Rts97] RTS Robotic Technology Systems LTD, Machine-vision for Industrial Manufacturing, August 1997,
            http://www.rtsfin.fi/english/machine-vision.html

[Shi85] Shimon Y. Nof, Handbook of Industrial Robotics, Wiley 1985, ISBN 0-471-89684-5.

[Mül92] Dipl.-Kfm. Ing.(grad.) S. Müller, Automation in the Press Shop with Industrial Robots, 23rd International
             Symposium on Industrial Robots, Barcelona 6th-9th October 1992.

[Oga93] M. Ogawa, The Work Handling Robot System for Press Machine, 24th International Symposium on
            Industrial Robots, Tokyo 4th-6th November 1993.
 
Back to Home Page