Definitions and Classifications

Let’s continue on with our definitions.  A kinematic chain is the term for all of the links and joints combined (kinematics is really just the study of motion).  A chain that is connected end over end is called a closed-loop chain, whereas a chain that has an end (such as most robots, and your arm) is an open-loop chain.  This isn’t too important to remember.  A mechanism is simply a device that consists of joints and links, and is connected to some fixed ground.  The mechanical part of the manipulator is a mechanism (i.e. the joints and links).  A machine (in this context) is a mechanism, but transforms energy into useful work using electricity, motors, microcontrollers, sensors, etc.

Robot manipulators can be classified in multiple ways – some of the ways that roboticists classify robots are by motion, by how many DOF it has (general purpose robots possess all 6 DOF, while redundant robots actually have more – which doesn’t seem possible, but it is, and it means the robot can be configured in more ways, to allow it to move around obstacles or in very tight workspaces), by the kinematic structure (open-loop, or parallel – which means that links are connected to every other link by two paths, adding structure, stiffness, but at the price of a larger robot and more complex mechanism), by the type of power (is it power by electricity?  or is it hydraulic?  or pneumatic? each offers advantages and disadvantages, but electric is most common), and by how the power is used (gears?  transmission system or direct drive?).  

A Robot Workspace

When you stand stationary and move you arm around, you can explore the outer boundaries of where you can reach.  This is limited by the degrees of freedoms of your joints, as well as the length of the links in your arm.  You can imagine you hand tracing out a 3D space in the air, which represents the regions you can reach while the rest of your body is stationary.  This is probably not something that we think of often, the outer boundaries (and inner boundaries) of where we can reach, probably mostly because we have the ability to move.  If you take away that ability, suddenly you’d become very aware of how far you could reach around you.  And for robots this concept is very important, since most robots (at least manipulator type robots) have a fixed base (non-mobile).  This 3D space is called the reachable workspace.  The reachable workspace is the limits of where the robot can reach in at least one orientation – a distinction which is important.  For example, when my arm is fully extended, I can only reach that space which my hand occupies with a single orientation (since all my links are fully outstretched).  If my arm was significantly longer, I might be able to reach that same spot, but with a different orientation of my arm.  This might be useful if there is something blocking my path, and so a straight extension will not allow me to reach the object I need to reach.  And so robots also have something called the dexterous workspace, which is the envelope that can be reached in any orientation by the end effector.  The dexterous workspace is much smaller than the reachable workspace (the dexterous workspace is a subset of the reachable workspace).