ADPD2140W

Data Sheet

Rev. 0 | Page 8 of 10

APPLICATIONS INFORMATION

GESTURE RECOGNITION

The unique angular response of the ADPD2140W coupled with

the high performance ambient light rejection of the ADPD1080

enables a robust and effective implementation of gesture

recognition. The following algorithm demonstrates recognition

of up, down, left, right, and click hand gestures based on data

from the four channels of the ADPD2140W:

1. Prior to operation of the ADPD2140W and the ADPD1080

for gesture recognition, calibrate the ADPD1080 clocks. See

the ADPD1080 data sheet for more information on how to

calibrate the 32 kHz and 32 MHz clocks.

2. Set the ADPD1080 mode of operation to sample mode by

writing 0x2 to Register 0x10, Bits[1:0].

3. Collect the data measured by the device. See the ADPD1080

data sheet for instructions on how to read data from registers

using the first in, first out (FIFO) and interrupts. Data is

available directly from data registers or from the 128-byte

FIFO in Register 0x60, Bits [15:0].

4. The data in the four output channels of the ADPD1080

with the following equations:

5. Prior to gesture event detection, offsets of the ADPD1080

must be digitally subtracted from each channel. These offsets

are not due to photodiode dark current and are set by the

ADPD1080 on-chip analog-to-digital converter (ADC).

Register 0x18, Register 0x19, Register 0x1A, and

Register 0x1B contain the ADC offsets for Timeslot A,

while Register 0x1A, Register 0x1B, Register 0x1E, and

Register 0x1F contain the ADC offsets for Timeslot B. The

nominal value for all offsets is 0x2000. To modify these offsets,

measure the 16-bit output of each channel, in ADC codes,

and add it to the existing 16-bit number in the ADC offset

register, SLOTx_CHx_OFFSET (nominally 0x2000). Then,

write to the ADC offset register with this result. When the

offsets are correctly subtracted, the intensity reading L is

close to zero codes with no objects in the sensor field of view.

6. The start of a gesture event can be defined as occurring

when intensity data crosses a preset threshold. Nominally,

this threshold must be set to 1000 codes. However, the

threshold can be adjusted to suit the application.

7. The end of a gesture event can then be defined as the number

of samples after which the intensity drops back below the

preset threshold, past a certain minimum number of samples

(nominally five samples).

8. Use the start and stop points of the gesture event to

determine whether the gesture was up, right, left, down, or

a click. For more detail on this process, see the following

pseudocode:

event = False

intensityThreshold = 1000 (should be

adjustable by the user)

clickThreshold = 0.07 (should be adjustable

by the user)

if event = True:

i += 1

if i >= 5 and L < intensityThreshold:

event = False

gestureStopX = x

gestureStopY = y

m = (gestureStartY –

gestureStopY)/(gestureStartX – gestureStopX

+ 1e-6)

d = sqrt((gestureStartX – gestureStopX)^2 +

(gestureStartY – gestureStopY)^2)

if d < clickThreshold:

gesture = ‘CLICK’

else:

if abs(m) > 1:

if gestureStartY >

gestureStopY:

gesture = ‘UP’

else:

gesture = ‘DOWN’

elif abs(m) < 1:

if gestureStartX >

gestureStopX:

gesture = ‘LEFT’

else:

gesture = ‘RIGHT’

else:

if L > intensityThreshold:

i = 0

event = True

gestureStartX = x

gestureStartY = y