Hand Gesture Recognition using PAJ7620 Sensor & Arduino

Overview
In this project, we will make Hand Gesture Recognition System using PAJ7620 Sensor & Arduino Board. This PAJ7620 gesture sensor can recognize nine gestures, including moving Up, Down, Left, Right, Forward, Backward, Clockwise, anti-clockwise, and wave. We can use the gesture recognition system in medical applications or control lights, robots, HMI, games, using IR LED and optical CMOS array.
Gesture Recognition
is a technology that uses sensors to read and interpret hand movements as commands. Gestures originate from any bodily motion or state but commonly originate from the face or hand. A gesture is an action that has to be seen by someone else and has to convey some piece of information. Gesture Recognition can be considered a way for a computer to understand human body language.
Earlier we learned about
APDS9960
Gesture Recognition & RGB Color Sensor. You can check the article to get started with Gesture Sensors.
Bill of Materials
S.N.
Components Name
Quantity
Purchase Links

Arduino Nano Board

Amazon
AliExpress

PAJ7620 Gesture Recognition Sensor

Amazon
AliExpress

16x2 I2C LCD Display

Amazon
AliExpress

Jumper Wires

Amazon
AliExpress

Breadboard

Amazon
AliExpress
PAJ7620 Gesture Recognition Sensor
This product is based on the PAJ7620U2 gesture recognition sensor, which recognizes gestures in 9 different directions, including up, down, left, right, front, back, clockwise, counterclockwise, and rocking.
The gesture recognition module uses the I2C interface and can be programmed and controlled by using the corresponding PAJ7620 Arduino library function. The signal returned by the gesture recognition module can be used as a control signal by the robot, thereby realizing the control of the robot.
The built-in recognition algorithm is quite intelligent, freeing both hands from blunt keys. Gesture recognition sensors can be used for non-contact control scenarios such as contactless mice, smart homes, car click device controls, robot interactions, and more.
PAJ7620U2 also offers built-in proximity detection for the purpose of sensing objects approaching or departing. The PAJ7620U2 is designed with great flexibility in a power-saving mechanism. The PAJ7620U2 is designed to operate from 2.8V to 3.3V over -40°C to +85°C and the pull-up voltage for the I2C bus and interrupt line is from 1.8V to 3.3V.
Refer to the datasheet for more information:
PAJ7620 Datasheet
Features of PAJ7620
Nine gesture recognition (Up / Down / Left / Right / Push / Pull / CW / CCW / Wave)
Gesture speed is 60°/s to 600°/s in Normal Mode and 60°/s to 1200°/s in Gaming Mode
The supply voltage is 2.8V- 3.3V and I/O voltage is 1.8V~3.3V
Working Current: 3mA-10mA
Detection distance: 50-100mm
Ambient light immunity: < 100k Lux
Built-in proximity detection
Flexible power-saving scheme
I2C interface up to 400 kbit/s, Pull-up voltage from 1.8V to 3.3V
Ambient light noise cancellation
PAJ7620 Pinout
PIN
SYMBOL
Description

VCC
3.3V/5V

GND
Ground

SDA
I2C data pin

SCL
I2C clock pin

INT
External interrupt pin
PAJ7620 Applications
Smart Home
Offices and teaching
Human-robot interaction
Gesture toys
Domatosensory game equipment
PAD Phone
Tablet Personal Computer
Automobile Application
Interfacing PAJ7620 Gesture Recognition Sensor with Arduino
Let us interface the PAJ7620 Gesture Recognition Sensor with Arduino using the simple I2C interface.
The connection is fairly simple. Connect the VCC, GND, SDA & SCL pin of PAJ7620 to Arduino 3.3V, GND, A4 & A5 respectively.
In my case, I am using an Arduino Nano board. You can use any other Arduino-compatible microcontroller board.
Source Code/Program
There are multiple libraries for interfacing the PAJ7620 Sensor with Arduino. One of the libraries is written by Seeed Studio. Download the library from the following link.
Download:
PAJ7620 Library
The following code is taken from the library example. Copy the following code and paste it to the Arduino IDE. Then upload the code.

#include <Wire.h>
#include "paj7620.h"
/*
Notice: When you want to recognize the Forward/Backward gestures, your gestures' reaction time must less than GES_ENTRY_TIME(0.8s).
You also can adjust the reaction time according to the actual circumstance.
*/
#define GES_REACTION_TIME 500 // You can adjust the reaction time according to the actual circumstance.
#define GES_ENTRY_TIME 800 // When you want to recognize the Forward/Backward gestures, your gestures' reaction time must less than GES_ENTRY_TIME(0.8s).
#define GES_QUIT_TIME 1000
void
setup
(
)
{
uint8_t
error
=

;
Serial
begin
(
9600
)
;
Serial
println
(
"\nPAJ7620U2 TEST DEMO: Recognize 9 gestures."
)
;
error
=
paj7620Init
(
)
;
// initialize Paj7620 registers
if
(
error
)
{
Serial
(
"INIT ERROR,CODE:"
)
;
Serial
println
(
error
)
;
}
else
{
Serial
println
(
"INIT OK"
)
;
}
Serial
println
(
"Please input your gestures:\n"
)
;
}
void
loop
(
)
{
uint8_t
data
=

data1
=

error
;
error
=
paj7620ReadReg
(
0x43

&
data
)
;
// Read Bank_0_Reg_0x43/0x44 for gesture result.
if
(
!
error
)
{
switch
(
data
)
{
// When different gestures be detected, the variable 'data' will be set to different values by paj7620ReadReg(0x43, 1, &data).
case
GES_RIGHT_FLAG
:
delay
(
GES_ENTRY_TIME
)
;
paj7620ReadReg
(
0x43

&
data
)
;
if
(
data
==
GES_FORWARD_FLAG
)
{
Serial
println
(
"Forward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
if
(
data
==
GES_BACKWARD_FLAG
)
{
Serial
println
(
"Backward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
{
Serial
println
(
"Right"
)
;
}
break
;
case
GES_LEFT_FLAG
:
delay
(
GES_ENTRY_TIME
)
;
paj7620ReadReg
(
0x43

&
data
)
;
if
(
data
==
GES_FORWARD_FLAG
)
{
Serial
println
(
"Forward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
if
(
data
==
GES_BACKWARD_FLAG
)
{
Serial
println
(
"Backward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
{
Serial
println
(
"Left"
)
;
}
break
;
case
GES_UP_FLAG
:
delay
(
GES_ENTRY_TIME
)
;
paj7620ReadReg
(
0x43

&
data
)
;
if
(
data
==
GES_FORWARD_FLAG
)
{
Serial
println
(
"Forward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
if
(
data
==
GES_BACKWARD_FLAG
)
{
Serial
println
(
"Backward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
{
Serial
println
(
"Up"
)
;
}
break
;
case
GES_DOWN_FLAG
:
delay
(
GES_ENTRY_TIME
)
;
paj7620ReadReg
(
0x43

&
data
)
;
if
(
data
==
GES_FORWARD_FLAG
)
{
Serial
println
(
"Forward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
if
(
data
==
GES_BACKWARD_FLAG
)
{
Serial
println
(
"Backward"
)
;
delay
(
GES_QUIT_TIME
)
;
}
else
{
Serial
println
(
"Down"
)
;
}
break
;
case
GES_FORWARD_FLAG
:
Serial
println
(
"Forward"
)
;
delay
(
GES_QUIT_TIME
)
;
break
;
case
GES_BACKWARD_FLAG
:
Serial
println
(
"Backward"
)
;
delay
(
GES_QUIT_TIME
)
;
break
;
case
GES_CLOCKWISE_FLAG
:
Serial
println
(
"Clockwise"
)
;
break
;
case
GES_COUNT_CLOCKWISE_FLAG
:
Serial
println
(
"anti-clockwise"
)
;
break
;
default
:
paj7620ReadReg
(
0x44

&
data1
)
;
if
(
data1
==
GES_WAVE_FLAG
)
{
Serial
println
(
"wave"
)
;
}
break
;
}
}
delay
(

)
;
}
After uploading the code, open the Serial Monitor. Then swipe your hand in different directions and make different gestures. The gesture indication will be shown in the Serial Monitor. The tests are straightforward, just plug in and start moving your hands.
There are two ways of detecting:
Near, which is on 5-15 cm distance with a 60° angle
Far, on a 15-30 cm distance with a 30° angle
You can modify the detection speed for all and for the forward and backward alone. Also, keep the module on a flat surface, otherwise, the clockwise and