Nuts & Volts - June 2007

code, the score is updated and the message “Score!” appears on the screen when the dot tracks the center of the target.

Figure 11 shows the single-axis output of a Memsic 2125 mounted horizontally on a stable surface. The 5,000 consecutive readings were collected at 10 Hz using a BASIC Stamp. The output of the accelerometer is Gaussian, with a mean of 4,908 microseconds and a standard deviation of two microseconds.

This output variability is significant in the game, in that the position of the white dot visibly jitters even when the wobble board is held stationary. I handle this jitter in accelerometer output with two simple filter routines written in SPIN. They take the form:

FIGURE 10. Memsic mounted in base of an OPTP wobble board.

if || (acc2 – acc1) > 10 ‘compare acceleration values
  acc1 := acc2          ‘update values if different
acc2 := memsic_output    ‘update acceleration value
acc1 := (acc + acc2)/2   ‘average acceleration

    t_squared := (target_x * target_x) + (target_y *
target_y)
    i_squared := i*i

if || (i_squared - t_squared) < 550
  if (timeRemaining >0)
    gr.text(-120,-80,string(“Score!”))
    totalPoints := totalPoints +1

The downside of this simple test for target and dot overlap is that there are multiple possible dot locations that will result in points. However, in practice, this simplification works well. A more complete and complex approach is to use a sine lookup table and compute the single point on the screen where target and player-controlled dot overlap in real time. Readers interested in following this approach can view several examples in the text accompanying the system.

The Memsic 2125, like every sensor, is imperfect. In addition to device-to-device differences in response and settling time, output from a stationary accelerometer varies randomly over time.

The filters are based on value differences instead of consecutive readings from the Memsic 2125 because the Propeller is operating at a significantly higher frequency than the accelerometer.

A final note regarding coding the Hydra is that this exergame relies on the Memsic 2125 object supplied by Parallax. This object or driver includes a self-calibration routine at startup that assumes the accelerometer is level. As such, the wobble board must be positioned close to horizontal at startup.

From Here

RESOURCES

1400

Hydra Game Development Kit, Memsic 2125 accelerometer, and software objects available from Parallax. www.parallax.com

1200

1000

FIGURE 11. Single-axis output of a level, stationary Memsic 2125; 5,000 consecutive samples at 10 Hz. Ambient temperature is 23°C.

Hydra Developer’s Forum. Forums.parallax.com/forums

Count

800

600

XGameStation. Andre’ LaMothe’s website for the Hydra, documentation, and forums.

400

200

Bergeron, B., Developing Serious Games. 2006: Thompson.

0

4895

4900

The exergame described here is a mere hint of what can be accomplished with the Hydra Game System. Consider adding a colorful background from the extensive library of image files on the CD-ROM that accompanies the system. Sound effects are another obvious addition to the exergame. More importantly, the wobble board interface can be used with very little modification on many of the sample games on the Hydra CD-ROM. Imagine playing a game similar to Break Out using a wobble board instead of a controller pad.

According to Andre’ LaMothe, developer of the Hydra and author of the text included with the system, a 512K expansion card is in production. The card will enable development of games with bitmapped graphics, more media, and, in general, greater complexity. Be sure to visit the forums on Parallax and XGameStation for any software