Before we start talking about robotic vision, it’s going to be very useful just to revisit some fundamentals of robots, and in particular, definitions of what a robot is.
For those of you who participated in the first online course, we presented this definition of what a robot is: It’s a goal-oriented machine that can sense, plan and act. And the important keywords are highlighted here in red.
A more casual definition that we introduced for a robot is that it’s a machine that can move—either itself or perhaps its hand—from place A to place B. So we might move its hand to pick up an object at place A, and put it down at place B; or it might movie its entire body. It might move along a corridor from place A in the corridor to place B in the corridor.
Let’s now look at some of the keywords in the first, more detailed definition, and one of the key concepts is sensing.
A robot fundamentally is a machine that can sense its environment. The sorts of things that it might want to sense is: where is the object that it needs to manipulate—where is the thing that it needs to pick up. Another thing that it might want to sense is where is the robot itself. A mobile robot in a corridor situation; whereabouts in the corridor is the robot? This is the classical ‘where am I?’ problem of mobile robotics.
Another thing a robot needs to do is to plan. So from its sensors it knows where it is; it knows where the things that it needs to work with are; it has got a goal, a thing that it wants to achieve. So it makes a plan of where to move from its current state to its goal state.
And the final part of the process is to carry out some action. Given that I’ve got a plan, now I have to carry out the step of the plan. I grab the object or I move myself from place A to place B.
Let’s recap what robots are, where they’ve come from and what they can do. Robots use sensors to understand their world and plan an action to achieve their goal.