There are four primary methods for programming a robot
Physical-setup , in which the operator sets up programs by physically fixing stops, setting switches, arraying wires, etc. This is characteristic of the simpler robots.
Lead-through , in which the operator leads the robot through the desired positions and locations by means of a remote teach box. (These points are recorded and used to generate the robot trajectory during operation.)
Walk-through , where the robot arm is physically manipulated through the desired motions (which are recorded and then played back by the robot-control during operation) .
Writing a software program , which is then executed when desired.
The emphasis in programming research today is on software programming of computer-controlled robots. Work on sensor-controlled manipulation is extending the scope for programmability. Interacting with the robot by means of software, provides more flexibility then the other programming methods, and allows for conditional actions or flexible adaptations. Various high level robot programming languages such as VAL (Unimation) and AML (IBM) are now beginning to become available to aid in the software generation.
Simons (1980, p. 107) notes that software programs can be divided into two types: "In explicit programming , the user requires explicit instructions for every action the robot must take. In world modelling , the robot is more a autonomousand can make decisions according to its knowledge. World-modelling systems, largely in the research stage, tend to require a considerable amount of computer power but are able to carry out complex tasks. The most important world-modelling languages are AL, AUTOPASS, and LANA..."