Trail sensor (state machine)

[vleeuwen]

During my preparation for the FtTxTrailSensor service (for MS-RDS), I made a description of the possible states and state transitions for this complex sensor.
This description contains two parts:
• The static part, the 4 states.
• The dynamic part with 12 state transitions.
In only two situation (out of 16), it is not possible to go back to the line.
See here.

The standard fischertechnik examples from the TX-C Trainings Labo, like the soccer_robot_x and trail_searcher_x are only using the static part. This limitation makes these algorithms sensitive to unexpected state transitions (v, w, x and y) and also in case of lost line (c,d).
In the fischertechnik examples, is in 1 out of 4 situations not possible to find back the line.
However with the use of two variables, these examples are extendable with a dynamic part in RoboPro.