In addition to the obstacle challenges, I’ve added a control screen to the GRIS program for light-follow mode. To get there, you click on the little flashlight icon at the top of the screen.

In case my artistic abilities aren’t clear enough: the screen portrays a pair of robot photo cells staring into a flashlight beam. The dark portion of the beam indicates that the robot is in non-triggered mode, where it searches (ie, turns), and the light portion of the beam indicates that the robot is in triggered mode where it moves toward the light.

(How the robot responds in non-triggered versus triggered mode was demonstrated in the video in yesterday’s blog entry, Grisbot Light Follow Test #1.)

Instead of turn and distance values, the scales at the bottom of the screen determine the photocell light levels at which the robot transitions between states. Note that the transition light level can be set at different points depending on whether you’re going from non-triggered to triggered state, or from triggered to non-triggered. This prevents ‘chatter’ or ‘bounce’ between states, and is much like how thermostats and relief valves are designed and like how microcontrollers are programmed to receive input from push buttons.

Why not just have fixed transition points? Well, because of changes in ambient light and the brightness of the flashlight beam, and also just to experiment (aka fool around) to see how the robot behaves differently as the set points vary.

I’m not satisfied with the way I’ve portrayed things here, and I’m open to suggestions on how to redesign this screen so that it can communicate its purpose to the user more clearly.

I’ve written all the underlying support code for this mode in the Scratch program. Next stop, I program the robot to respond to the flashed data. And I hope that I can get that done for the Seattle Robotics Society Meeting on Saturday.