Complexity of tasks for Robots

Robotic applications can be broken down into four different levels of complexity depending on the amount of sensing and decision making required. The applications that fall into the first category are the easiest to automate and those in the fourth category the most difficult and in many cases not possible with current technology.

Level 1

Applications that can be achieved using a simple robot using jigs and fixtures to position components and tooling to achieve the required accuracy. Example applications include spot welding, adhesive or sealant application, painting etc. In order for these to work the variation of the components along with the repeatability of the robot used must be within the tolerance required by the process.

Level 2

Applications requiring sensory feedback in order for small modifications to be made to the program to account for variation in the components. Examples for this include arc welding, automotive window glazing and spare wheel mounting. For arc welding the accuracy with which thick section metal sheet can be cut and formed lies outside of the tolerance in torch position for quality arc welding. To overcome this sensors are often used to locate the position of joints and seam tracking devices which can use either vision systems or through arc sensing are used to track the seam once it has been located. Seam tracking is essential for most applications of arc welding as the localised heating caused by the process tends to cause the component to warp.

Level 3

These applications require more complex sensory capabilities such as pattern recognition. They also tend to require complex decision making based on this feed back. Examples where this level of application has been automated are few and tend to be in the assembly area. A number of systems now exist for mounting wheels onto car bodies. This requires the robot to find out where the wheel hub is and in what orientation it is such that the holes in the wheel can be matched to the studs. It then needs to work out the orientation of the wheel in the rack and calculate how to pick it up so that it can place it accurately onto the hub. It then runs the wheel nuts onto the studs to the correct torque level.

Level 4

The most difficult applications are those involving unpredictable behaviour of either the components or other equipment within the cell. Operations such as handling of flexible components i.e. carpets or water hoses for example are examples of this. In the future robots may also link more intelligently with humans so that they can judge for themselves when it is safe to operate without having comprehensive guarding and safety interlocks everywhere. Some robots are already being developed for the nuclear industry that have tactile sensors covering the arm such that collisions can be detected before any damage is caused.