All robot types are made up of a series of joints that can be likened to human limbs. Some are simple such as the Cartesian co-ordinate but others are much more complex, they can replicate human arm movements in most cases and with the jointed arm type they can even exceed what can be achieved by us. With some of the jointed arm robots a greater range of movement is achieved by adding 360° swivel joints normally at the wrist joint which allows even greater reach capability. The joints are so similar to ours that they are even named shoulder, elbow and wrist. Robots, when ordered from the manufacturer robots are not normally supplied with grippers or end tools (normally referred to as end-effectors), this is because each robot will be used for a unique and specific application by each customer. However in some cases they can be supplied to order to suit specific applications. The word robot is derived from the Czechoslovakian word for slave.
Over the next few pages I have listed the main types of robot, their applications, their advantages and limitations.
TRICEPT & HEXAPOD ROBOTS.
Fig 0.1
The tricept and hexapod type robots normally use linear motors to control the angle of the end effector. The tricept robot as shown in fig 0.1 uses three of these legs in conjunction with a central pillar to hold the head in position and then has a standard wrist joint to achieve the required orientation. A hexapod uses six legs and achieves both position and orientation in the same way. Both of these robot structures are very rigid but both have the disadvantage of small working envelopes and limited orientation ability. These structures are generally used for machining operations where machine tool level tolerances are not required but greater flexibility would be needed.http://www.robotmatrix.org/images/TriceptRobot.gif
Fig 0.2CARTESIAN CO-ORDINATE ROBOTS. Work Space Envelop of the Cartesian Roboti
A common type of robot in industry is the Cartesian co-ordinate robot; it has a very high rigidity due to its strong external frame. It can therefore often be seen in machine tools and co-ordinate measuring machines as there is very little deviation due to flexing of the frame. It is widely used by many manufacturing and packaging companies as pick and place robots which is a result of its ability to repeat operations with a high degree of accuracy. This is a simple type of robot that has a working envelope that is surrounded by its frame as shown in fig 0.2, normally it is square or rectangular in shape and this type is used widely through the Ford motor company in Bridgend plant. The one disadvantage of this type of robot is that the linear joints can be difficult to seal which makes them prone to ingress when used in damp and dusty environments.
CYLINDRICAL CO-ORDINATE ROBOTS.
Fig 0.3As with the Cartesian robot the cylindrical co-ordinate are similar in that they have a high degree of rigidity and are therefore most suited for jobs requiring straight line moves. They are relatively simple to program as their motion path is easy to visualise. They are well suited for applications that require reaching into cavities with an extendable probe. The disadvantage of these robots is their inability to reach around objects and the amount of clearance required behind the robot. The linear joint as with the Cartesian makes them unsuitable for working in dusty or damp environments. Fig 0.3 shows a cylindrical co-ordinate robot. http://www.denso.com.au/var/denso/storage/images/products/aftermarket_industrial_products/robotics/product_range_overview/4_axis/cs/2101-3-eng-AU/cs_large.jpg
POLAR CO-ORDINATE ROBOTS.
Fig 0.4Polar co-ordinate robots were the first types to be used in any type of industrial application. The reason for this is that they lent themselves ideally to the hydraulic drives which as shown in fig 0.4 were the main type of drive at the time. Since the electrically powered drives have become more and more advanced and sophisticated this structure have been replaced in the main by jointed arm robot types, although hydraulic robots are still seen in industry today mainly performing the more simple tasks such as spot welding etc.Polar Robot
SCARA (Selective Compliance Assembly Robot Arms).
Fig 0.5'SCARA' robots are designed to be used specifically in peg board assembly applications and are widely used in the electronics industry but are also used in many other types of industry for relatively small components e.g. placing valve seals in Ford's engine plant at Bridgend. The SCARA type robots are very stiff in the vertical axis but have a degree of error in the horizontal plane which allows for minor errors in placement of components. Due to their small physical size they have a very high speed and accuracy making them ideally suited for use in assembly, palletisation and machine loading as shown in fig 0.5.http://www.intelligentactuator.com/images/scaragripper.jpg
JOINTED ARM ROBOT.
Fig 0.6The jointed arm type robot is the most versatile of all robots due to the fact that it has so many axis movements. The benefit of these axis movements is that the robot can achieve an infinite amount of positions within its working envelope. This robot type most closely resembles the human arm in that it has a shoulder rotation, an elbow rotation, a wrist rotation and with the base rotation mimicking the human hip rotation it allows for 270Ëš rotation in most cases but in the more sophisticated ones 360Ëš degree of work is achievable. The downside of this however is that because of its many axis movements it can be a very difficult to program as it is difficult to visualise all of its axis movements during programming. As stated before it has a very wide range of applications such as paint spraying, arc and spot welding this type of robot is suitable for specific tasks which require either a high level of manoeuvrability or requiring a high level of accuracy with regards to axis movement. In addition to the high level of manoeuvrability these types of robot can be mounted to gantry's or Cartesian robots giving it a much larger envelope as shown in fig 0.6.http://www.kuka.com/nl_media/06/bild_kuka_jet.jpg
WORKING ENVELOPE.http://63.234.227.130/dts/osta/otm/otm_iv/otm_iv_4fig06.gif
Fig 0.7A robot's working envelope is the range from which it moves around, the area of its movement or limits. It is the shape created when a manipulator reaches forward, backward, up and down. These distances are determined by the length of a robot's arm and the design of its axis. Each axis contributes its own range of motion. A robot can only perform within the area of this work envelope. Still, many of the robots are designed with considerable flexibility. Some have the ability to reach behind them as shown in fig 0.7, the jointed arm with its complex joints has a very agile working envelope and when coupled with a gantry enables the agility to transfer to a much longer area. Gantry robots defy traditional constraints of work envelopes. They move along track systems to create large work spaces. It is important to take work envelope into consideration when choosing a robot. You should try to make your work envelope match your application. Different work envelope capacities will suit different environments and needs. What are the requirements of the job it would have to perform? Do you need a robot to handle a large part, or traverse a long distance? Then choose a model with a larger work envelope. Consider the advantages of a gantry system. Is your work in a contained space or with tiny objects? Then select a robot with a smaller work envelope.
GRIPPERS, END-EFFECTORS AND MANIPULATORS.
From the above descriptions of the 6 main types of robot structures it is shown that they all have more than one type of application the way in which we can obtain multiple applications is by changing the end effectors on these robots. If you were to relate to the human body where the arm is the equivalent of a jointed arm robot then the end effector could be a pen, fork or screwdriver etc. There are many types of end effectors which can be used with robots such as grippers for pick and place operations, drills, taps, welders and the list goes on. First of all before you can decide which type of gripper to use you need to consider the following factors;
Gripping force.
Weight of the object to be lifted.
Size or shape of the object.
Speed of movement.
Operating restrictions.
SURROUNDING ENVIRONMENTS & SAFETY REQUIREMENTS.
The primary consideration when designing the environment of any robotic application is safety. When designing and installing any robotic application the safety of the operators, maintenance personnel and anyone working in the vicinity must be the primary concern. The main hazards associated with modern robots is that they operate at very high speeds, with extreme power and provide very little warning as they use mainly electric drives which are nearly silent in operation. Therefore it is essential to ensure that there are sufficient precautions in place to control or ideally eliminate the possibility of accidents and/or injuries occurring as a result. The main way of safeguarding employees is to use sufficient guarding and enclosures in order to physically separate them from the machinery. As well as guarding there must be sufficient power control/isolation facilities in placed for maintenance personnel when undertaking tasks within robot enclosures (ECPL, electrical control power lock-off). Another consideration is the safety of the programmer when using the teach method in which case a dead man's switch is normally used. This method works by allowing control of the robot only while the operator holds a safety button on the teach pendant, as soon as the button is released the robot freezes instantly. Control can only be regained by depressing the safety button on the teach pendant.
SOME OF THE MOST COMMON SAFEGUARDS ARE DESCRIBED BELOW.
PERIMETER FENCING. http://www.airoil.com/uploads/assets/product_images/frame_world-robot_guarding.jpg
Fig 0.8This type of guarding is produced from rigid panels, usually mesh which can be permanently fixed to floors and other suitable anchor points but must be removable for the purpose of maintenance. Normally these panels/guards are removed using keys which in the case of the engine plant Bridgend are 8mm/6mm Allen Keys although control panels are normally accessed with a specialist key that is not easily accessed by non trades personnel. It is important that the type of guards used allow a full view of the robot in its operation while material chosen must still offer sufficient protection to the operators and maintenance personnel. The most widely used material used throughout industry today as robot guarding is steel mesh as it allows adequate protection against personnel straying into the robot's path/envelope and also allows for ideal viewing but unfortunately would not stop projectile hazards and hazards associated with welding (Flashes & radiation) and airborne hazards like paint spray fumes etc. Other controls could be used /added in these cases such as welding curtains or Perspex in order to contain the hazards. All enclosures must be provided with a recommended access point that is only accessible by disarming the appropriate safety controls like 'Safeguard' and 'Armguard' interlocking mechanisms
INTERLOCKING DEVICES.
Guard operated interlocking devices require keys in order to allow access through gates in the guarding. Normally keys are only accessible via the area foreman by completing the relevant documentation. Also commonly in use throughout the engine plant are locking devices that require the transfer of keys from a control box to access gates.
LIGHT CURTAINS.
Light curtains are used in areas where loading of components is required normally by forklift trucks where the light curtain is obstructed then the robot requires re-setting by the truck operator in order to allow the robot to continue. This ensures that the operator is at a safe distance and work is safe to continue. The light curtains operate by detecting anything that breaks the light beam between the transmitter and the receiver producing an output signal to the robots control unit. This method provides instant access to the workspace but it provides no protection from any projectile hazards from the robot.
PRESSURE SENSITIVE MATS.
When using pressure sensitive mats you must consider some key calculations in order to provide an environment that has relative safety; length of stride, speed of approach and the systems response time all require consideration when deciding on the mat's placement but as with the light curtains no protection is provided using this method alone.
TWO HAND CONTROL DEVICES.
As is the title suggests two hands are required to use this hold-to-run control ensuring both hands are clear of moving parts during operation. The robot is designed to stop when both buttons are released and changes cannot be made to the robots operating system changed while the buttons are held down.
PRESSURE SENSITIVE EDGES.
Flexible edging strips which can be fixed to the edge of a moving part such as a machine table or powered door where there is a risk of a crushing or shearing hazard. Another common use of these strips are used on raise/lower tables where they detect any object that can be trapped below when lowered such as feet, it works by producing an output to the control unit when an object is detected by coming into contact with the strip.
POWER UNITS ON MODERN ROBOTS.
Industrial robots are powered by three types of drive systems, they are electric, pneumatic and hydraulic. Each of the power sources have their own benefits and characteristics that make them more suitable for specific applications. Electric motors are efficient, require little maintenance, and are relatively quiet in operation. Pneumatic robots use compressed air and therefore require a separate source of energy(compressor/propane cylinder)and come in a wide variety of sizes.. Hydraulic robots use oil under pressure are best suited to applications where strength is a major benefit although this type tends to be noisy, quite large and much heavier than other power sources. As with the pneumatic robots the hydraulic types also need an auxiliary source of energy to move the fluids through its components via Pneumatic and hydraulic tubes, fittings and hoses that connect the components and distribute the energy. Here are the three options;
Pneumatics
Hydraulics
Electrical (Motors)
The above three power sources are briefly described below stating their advantages and disadvantages and some maintenance requirements that should be considered when carrying out maintenance tasks upon them.
Pneumatics.
Although this power source is widely used ion applications throughout the manufacturing industry isn't generally used as a main power source for robots. The main reason for this is due to the fact that it does not possess the required power capabilities and accuracy, although pneumatic end effectors are widely used due to the fact that they are readily available and cheap, they are also fast in operation allowing for quick cycle times. Pneumatics is a relatively clean source of power which lends it to applications where cleanliness in paramount like food industry and pharmaceuticals. The main disadvantage of pneumatics are they are not as powerful as hydraulic units using the same size equipment and would require much larger, more expensive pneumatic system. Also pneumatics are not controllable as hydraulics because the gas/compressed air used are more easily compressed than hydraulic fluids making positioning an little trickier. The other restriction when using pneumatics in robotics is that a separate reservoir must be provided in order to store compressed gases which can make them bulky. It is also necessary to ensure the air is clean in the system therefore filters (F.R.L.'s) are required and must be regularly maintained as they wear quickly. The robot drives are usually a combination powered by pneumatic, hydraulic, or electrical power types of energy, and the selection is usually based upon application requirements. For example, pneumatic power (low-pressure air) is used generally for low weight bearing robots.
Hydraulics.
Hydraulics is the power source that was the most widely used power source for early robots because of its advantages over pneumatics and they are still used widely today in applications where power is the main requirement as they are more powerful than the electrical power drives. Hydraulics are generally used in industry where power is needed but speed is not necessary and where a clean environment is not essential. The disadvantages of hydraulics are that they tend to be very expensive not only to install but also to maintain, due to the cost of the components and the hydraulic oil. Hydraulics systems are very dirty and can cause a slippery environment due to oil leaks, this leads to requirement for the regular maintenance of seals and hoses. Due to the leaky nature of the hydraulic systems regular cleaning will have to be adopted in order to minimize the risks of slips trips and falls etc. Filters are required for cleaning the oil continuously and to control the system environment. Hoses should be checked at regular intervals to ensure no damage is caused by blown pipes due to the high pressures involved. Therefore we can say that hydraulics as a power source for robots are mainly suitable for applications where power, controllability and a smooth action are major factors to consider when designing a robot.
DC Servo Motor
Dc servo motors use variable dc voltages as a form of speed control, a tachometer driven by the motor provides speed feed back to a control system called a servo loop and a rotational position sensor provides position feedback to form a self contained positioning system. The servo motors are used in applications where precise position control and high speed operation is a necessary characteristic. The dc servo motor can develop a large amount of torque, but care must be taken that it is not overdriven as the permanent magnets can be demagnetised. A substantial amount of control circuitry is required for a motor of this type, so it usually is not used unless its specific operating characteristics are required.
AC Synchronous Motor
These types of drive are best suited for applications that require continuous rotation, this is especially true where a constant speed is required. Speed control is dependant on the frequency of the supply voltage. Three phase synchronous motors are used in applications where higher torque is required and typically it is rated at more than one horsepower. It can be used in proportional speed control applications using circuitry in order to change frequency and relative phase. The starting torque is poor in single phase synchronous motors, compared to other types of motors.
Stepper Motor.
When stepper motors are used with external circuitry, direct digital control is achievable. Due to its field windings, the stepper motor operates in a series of steps rather than a continuous motion. Very precise control can be achieved by using these steps along with the necessary gearing to achieve almost infinite control. This is an important requirement of robots that use indexed movement. One limitation of a stepping motor is that torque is inversely proportional to speed. They are advantageous in applications where you need to control;
rotation angle
speed
position
and synchronism
Stepper motors are derived from printers, plotters and hard disk drives etc but have made the jump to robotics due to their precise controllability.
MAINTENANCE REQUIREMENTS.
As with all machinery that has many moving parts there is a requirement for some routine maintenance in order to maintain a fully functioning and safe working robot. The best type of maintenance in robotics as with any mechanical system is preventive maintenance. When carried out on a routine basis manufacturing companies will make huge savings by reducing down time and loss of production. By using a preventative maintenance schedule, the robotic equipment in the work place should last a lot longer and if planned correctly can be carried out during non productive times with minimum disruption to the plant's output. Failure to properly maintain various parts of a robot will almost certainly result in failure of components which is likely to result in damage to the robot, its surroundings or in extreme cases injure or kill personnel. Well designed preventative schedules will include daily checks which may be mainly visual or will include lubricant top-ups etc.
Specific parts may need to be inspected at specific cycle times as indicated by the manufacturer. Lubricants should be replaced on a consistent basis, and terminals, connectors, harnesses and brake leads, tested. Robots have many articulating joints that require proper lubrication in the form of gear oils and special grease. All maintenance personnel who perform work on robots should be well trained and understand the complete workings of the machine. They should also be trained in proper lock out tag out procedures. Not only will this save destruction to property but aid in the down time due to injury that may be caused by improper shutdown of the robot.
FEEDBACK LOOPS.
People use a range of senses in order make sense of our environment which we could call feedback, for example if we need to move an object from one place to another we call upon our senses to know how much force is required to push or lift it. As we already know we use five basic senses;
Taste
Sight
Hearing
Touch
Smell.
The messages that are returned from our senses are processed by our brain, similarly on a robotic machine the messages are sent from the robots sensors back to a computer that will decipher what the next course of action will be. on a computer a which decides what we need to do with it. Robots need to be able to detect what is going on around them and react to the information that is returned to them. But in the case of both human and robots, the way the information from the senses & sensors is processed can have a huge effect on what course of action is chosen. There are two possibilities;
Closed loop systems
Open loop systems
Closed Loop systems.
Below is a diagram of a central heating control loop system. The set point is where the desired temperature is set then the thermostat controller controls the release of fuel to the furnace fuel valve. The heat produced is then lost through the home heating process(radiators) and a temperature sensor in the home feeds back temperature information to the set point, if the desired temperature is met then the control loop is stopped until such time as the temperature sensor falls below the desired set point. Open Loop control systems.
Fig 0.9
FIGURE 0.2Open loop control systems are different to closed loop systems as they don't have any form of feedback through sensors. This means that they are unable to regulate the output. The diagram below (Fig 0.9) shows an open loop system that is used to heat a swimming pool using solar energy. You can see that the pool water is drawn out of the pool using the pump, pushed through the filter and into the solar collectors (panels). Heat is then absorbed by the water and passed through a secondary heater after which it is deposited back into the pool. As there is no thermostatic device in the loop there is no automatic heat control.
Closed Loop.
ADVANTAGES.
Wide range of applications
Maximum efficiency
Greater stability
Once setup it will require less human input
DISADVANTAGES.
More expensive to introduce
More complex system
OPEN LOOP.
ADVANTAGES.
Simple design
Suitable for basic systems
Easy maintenance
DIADVANTGES.
Not suitable for complex systems
Inefficient
No feedback from the output
DRIVE RATIOS.
The drive ratio is the term given to the relationship between the drive of the motor and the driven component of the robot. The motor and components may be joined by toothed gears or belt & pulleys etc, each of these types have their own characteristics which make them suitable for a specific operation. The drive ratio can work in two ways;
To step-up speed
To step- down speed
To step up the speed of a robotic application the drive gear must be increased in comparison to the driven one, therefore in a system where the drive gear has 32 teeth and the driven gear has 8 then for every 1 turn of the drive gear the driven one will turn 4 times. In this case the speed is increased but the disadvantage is that there would be a reduction in torque and reduced controllability. Alternatively when stepped down, the speed of the driven gear is decreased therefore the motor speed must be increased in order to compensate and as a result the torque increases along with the controllability and smoothness of motion. The gears of a robot work on the principle of Torque v Rotational velocity/speed where torque is inversely proportional to speed. The bigger the difference in gear size the bigger the difference in speed and torque.
Types of Sensor and Feed-back facilities.
Every robot requires sensors making it very advanced when compared to standard machinery. The benefit of these sensors is feed-back, which is the term for the information that is returned from all of the systems sensors to help it make sense of its environment. When designing and programming robotic systems it is important to have sufficient foresight to place the sensory equipment in the best possible positions which will enable it to understand what position each of its components are in and make sense of its surroundings. Sensors work by providing feedback to the control system about the position of actuators or other components, how far they have moved, the speed and acceleration etc. Some sensors can provide information of pressures and forces. Some advanced sensors such as optical sensor, thermal sensors and vision sensors which can create more vivid images of the surroundings and provide more detailed information to the process control system.
Accuracy.
Fig 1.0In order to perform tasks where accuracy is the main characteristic quality, devices are used in order to measure movement with minute precision. These devices are called 'encoders' and are generally added to the end of the motor shaft which is not connected to the robot arm. By doing this you can gain a high degree of accuracy because further control of the robots positioning can be gained by a robotic arm's movement. If an encoder's disc has 360 steps, then for every 1 step of the encoder we would get a movement of one degree and by moving 180 step counts the robot arm will move 180 degrees. Most encoders however have a large number of steps, on a disk with 3600 steps for every 360 degrees of motion, 90 step counts will only move the robot arm by 9 degrees and therefore we gain a much higher degree of accuracy. encoder
"Another simple way of obtaining high accuracy and speed in robotic cells is by utilising the ratios of motors and their encoders. As we know every motor has 360 degrees of motion if we relate this to the robot are then if the motor is connected directly to the robot arm then it too will have 360 degrees of motion. The way in which we then control this motion is by adding and encoder onto the end of the motor this measures the motor position by turning its 360 degrees of motion into a series of steps which it measures. The way in which it does this is by usually using a plastic disc with notches cut out of it and a light source. The diagram that is shown below is from a website and it best shows how the optical encoder works" http://www.brighthub.com/engineering/mechanical/articles/26214
Perceived intelligence of robotic systems.
Robots can on times be 'perceived' to be working under their own control, but are they displaying any intelligent thinking? Thanks to their sensing equipment they are able to make some sense of their surroundings and make decisions about how to operate within its own environment. Robots can, if they are programmed to do so learn from its experiences by bumping into things, remembering its co-ordinates and avoiding them the next time. As humans we too can absorb this sort of information because we store it in much the same way as robots by using our 'memory' or with robots 'memory banks'. But that is where the similarities with humans end and that is because after robots have stored information they can only carry out tasks which they have been pre-programmed to carry out. In order to carry out any tasks robots must ultimately rely on external input from humans in the form of programming and software which mean that they are completely reliant upon us. As the name's origins suggest, 'robot' is derived from the Czech word for slave. The one thing that robots will not be able to recreate in the foreseeable future is our consciousness and ability to think freely for ourselves. So using current technology and the programming and software currently available we can say that robots are not capable of their own intelligence but can understand why they sometimes are 'perceived' as being so by so many people.
Programming .
There are many different methods of programming the various robotic cells and their peripheral equipment these range from task programming, using a teach pendent, having a manual data input or by using peripheral equipment on the robotic cell. All of these methods have their own advantages and their own disadvantages but all are very viable methods used to program any robotic cells.
Fig 1.1The first method is task programming which is also called the 'lead through' method of programming. This was at first the main method of programming robots but as the other methods have advanced this type of programming has decreased in use but it is still used for a number of various applications. The way this method works is by an operator actually moving a robot through its movements and actions. This method lends itself to applications such as paint spraying on cars and other components but it is very unreliable as it will perform the exact same movements every single time regardless of any change to the conditions. Therefore if you want to paint a different model car for example then you would have to completely reprogram the entire robotic cell for this new car and then reprogram again if you wanted to change it back to the original model. This is also a safety issue because once it has started it will just perform its movements and actions regardless of whether there is a part in the system or not. The reason it is widely used with car spraying though is that by leading it through by hand the operator can position the robot easily into the harder to reach places on the component this would be very difficult to do with the other methods.
Teach Pendant.
One of the most widely used method to program robots now is by using a teach pendent as shown on previous page in fig 1.1. This method involves trained personnel using a device known as a teach pendant to program the robots positions. This method is commonly used with co-ordinate Cartesian robots also known as gantry robots because it is easier to see their movements and then program the position with the pendent. The advantage of this method is that it is a very easy way to program because you can actually watch the robot going through its motions so you can actually see what it is going to do. The disadvantage however is that if there are any obstacles in the robots way then it can be very hard to visualise how to program the robot to go around it.
Manual data input.
The other method which is used is the manual data input method it is also known as live programming as the robot will have already been installed and will have been running previously. This method involves linking a computer to the robotic control system when the robot is running and then altering the program from there using the computer. The advantage of this method is that you can watch the robot running and then see your alterations in effect straight away. However the disadvantage is that while you are able to change the program easily you won't be able to test the changes effects before running the system you would have to just have to trust that the changes made are accurate and correct. For this reason only the people who have been trained and are skilled with the robotic cell should attempt this method.
Peripheral Programming.
The other method used to program robotics is by using the peripheral programming method. This method is very similar to the way CAD systems are used to generate NC programs for milling machines because a robotic program can also be created from a CAD program. This is usually done by an outside source and is then sent to the place where the robot is and then installed to the robot. The obvious disadvantage of this is that the programmer can actually work with the robot and can only go on what the CAD program is telling him or her. This can lead to problems when the program is installed to the robot because the programmer may not have accounted for obstacles which the robot might meet. This method has only recently begun to be used but it does have its own advantages;
Reduced down time for programming.
Programming tools make programming easier.
Enables concurrent engineering and reduces product lead time.
Assists cell design and allows process optimisation.
BIBLIOGRAPHY.
www.osha.gov/dts/osta/otm/otm_iv/otm_iv_4.html
www.Festco.com
www.manufacturingtalk.com
www.torontosurplus.com/ind/DATA1056.JPG
www.youclaim.co.uk/robot-fatality-and-accident-at-work-cases.htm
www.howstuffworks.com/jobs.htm
www.wikipedia.com/industrial robot
www.abb.com/robotics
www.bukisa.com/.../138337_actuators-of-industrial-robots-motors
www.channel4.com/science/microsites/R/robots/s_process.html page 22 www.brighthub.com/engineering/mechanical/articles/26214
