THE VIRTUAL WALL UNIT

Last modified:

INTRODUCTION

Scooba's Virtual Wall Unit, (VWU), uses nearly the same package configuration as the Discovery-Scheduler's VWU, but Scooba's has the well-known, simple, user-interface, since it needs no programming interface. At present, we assume Scooba's VWU uses the same circuit-schematic and components as the older-style VWUs which accompanied Discovery-model purchases. With that said, there is not much else to say about the VWU in this section! However, different surface treatments on the "collimating" tubes were noticed, blackened on the old, reflective on the new, and that triggered a little study of Scooba's radiance-pattern in a nearly horizontal plane.

To save readers the drudgery of reading further, here is a little summary of that study. Yes, the reflective walls produce definite off-axis IR-intensity lobes, about the central-axis of the horizontal wall beam; and NO, the impact of those artifacts will not necessarily affect Scooba in a manner different than the way Roombas respond to the emittance pattern produced by the original-style VWUs.

Package Configuration

Within the series of VWUs issued with Roombas, there was a fair amount of empty space. Cells, were at the unit's bottom, while electronic assemblies were at the top; and there was little in between. That design resulted in a stable platform, but was considered difficult to grasp and lift using finger-tips.

Those characteristics have been discarded in the Scooba-VWU. Cells stand up on end, which allowed slimming the width of the package, thus easing the pick-up difficulty. A sliding, side-hatch could then be added so D-cells replacement would not require a tool to accomplish.

VWU-controls were moved off the top surface and relocated on a sloping face below the horizontal output beam. Figure 1 shows the upper portion (roughly the upper half) of a Scooba-VWU. That image is the left-most scene in the panorama of images provided in Figure 2 -- which will be explained momentarily.

In Figure 1, the VWU is powered, and the unit has been rotated (around a vertical axis) about 20-degrees off the camera's LOS (line of sight). At this angle, and larger angles, the robot would not be repelled. But notice, the port is still scattering some IR-radiance. This particular image does not emphasize it, but when the output port is viewed with the visual-response of the eye, the port's collimating tube appears to be metal-plated; in other words, it reflects light; which is opposite to the desired effect.

Goniometry

That reflective condition suggested there should be peaks in IR output, just to each side of the central maximum; and, that possibility caused a rough set of "goniometric" measurements to be done. In this case, goniometry involved making rough intensity measurements, via a 'fixed-iris' digital-camera, at "zero- degrees" and at several off-axis angles. After the central image was captured, the VWU was indexed in azimuth just enough (plus and minus) to search out a dark or dim condition -- which was then photographed. Moving to larger azimuth angles, intense glints from the tube-wall were found and captured. Finally, the far-off-axis, dark-onset positions were captured. The result is shown as the panorama of images in Figure 2 (Hint: The feature to study is the circular port through which the horizontal-beam-LED outputs its energy.).

To rotate the VWU, it was placed on a small rotary table (a machine-shop tool). The zero-degree LOS was first established by rotating the VWU in azimuth, back and forth, while the camera was incrementally indexed in azimuth and elevation pointing angles, and laterally (up/down) to optimize pointing along the LED-source's maximum intensity axis. The center image shows that condition. The three images to the left, and three to the right, were obtained as follows:

1. While watching the LED's output through the camera, the rotary-table was rotated away from zero (say, first in a positive sense, then negative) until the LED's intensity fell to a minimum, for each case. Snaps at +9 and -6.5 degrees were captured (i.e., the pix just left and right, respectively, of center).
2. Moving to larger angles, of both signs, to peak the reflected (i.e., glint) IR intensities, snaps were made at +12 and -10 degrees, thus providing the bright-port pix just inboard of those at the very ends of Figure 2.
3. Finally, the table was rotated to the two angles where the LED's output fell to a useless minimum. Those angles are +20 and -17 degrees, for the end pix.

Remarking about (1), Scooba might travel into these dark zones, but an owner may never notice, because the bot will quickly maneuver an exit once it encounters a bright lobe of intensity at either side.

Then, about element (2), owners should expect the robot to be repelled by these high-level, glint-intensities in the same manner that Roombas work with the large beam-angles from their VWUs.

Item-3 pictures represent the cut-off angles, after which no more glints are produced.

With the minimal effort so far expended, it is not possible to claim that Scooba will show any difference in interacting with its VWU, than Roombas do with theirs. To be definitive, a more capable goniometer would have to be used to compare spatial-radiance from VWUs having black vs. reflective collimating tubes, while also using a Roomba/Scooba-style top-optic and its photo-detector as the profiling sensor (rather than a camera). A fun project, but one that is not scheduled.

If you wish to return to the main (Scooba Technical Comments) document, click Main

Otherwise you may return by pressing the 'Backspace' key.

This page is currently maintained by G. Plews