Arduino UNO एक R4 WiFi based AI/ML Gesture Recognition with ADXL345

In this step-by-step tutorial, you’ll learn how to use Edge Impulse, a leading platform for embedded machine learning, to build a gesture recognition project. By the end, you’ll know how to collect data, train a custom ML model, deploy it to your Arduino UNO R4 WiFi, and run real-time AI inference directly on the hardware

Sep 3, 2025

•

7812 views

•

2 respects

•

GPL3+

Devices & Components

SmartElex Bharat AI Innovators Kit Powered by Arduino

Software & Tools

Edge Impulse Studio

Arduino IDE

Project description

Code

Final Gesture Recognition Code

cpp

Ready-to-run inference sketch

1#include <your_project_inference.h> (the exact name is shown in the Arduino
2examples).
3#include <Wire.h>
4#include <Adafruit_ADXL345_Unified.h>
5#include <your_project_inference.h> // <-- replace with your EI library header
6Adafruit_ADXL345_Unified accel = Adafruit_ADXL345_Unified(12345);
7// EI expects data every EI_CLASSIFIER_INTERVAL_MS
8static bool debug_nn = false;
9void setup() {
10Serial.begin(115200);
11while (!Serial) {}
12Wire.begin();
13delay(50);
14if (!accel.begin()) {
15Serial.println("ERROR: ADXL345 not detected");
16while (1) { delay(10); }
17}
18accel.setRange(ADXL345_RANGE_4_G);
19if (EI_CLASSIFIER_RAW_SAMPLES_PER_FRAME != 3) {
20ei_printf("ERR: Model expects 3 axes\n");
21while (1) { delay(10); }
22}
23Serial.println("Ready for inference (move the board: up/down/circle).");
24}
25void loop() {
26// Capture one frame of data (features window)
27float buffer[EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE] = {0};
28for (size_t ix = 0; ix < EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE; ix += 3) {
29uint64_t next_tick = micros() + (EI_CLASSIFIER_INTERVAL_MS * 1000);
30sensors_event_t e;
31accel.getEvent(&e);
32buffer[ix + 0] = e.acceleration.x; // m/s^2
33buffer[ix + 1] = e.acceleration.y;
34buffer[ix + 2] = e.acceleration.z;
35// keep precise timing
36int32_t wait = (int32_t)(next_tick - micros());
37if (wait > 0) delayMicroseconds(wait);
38}
39// Create EI signal
40signal_t signal;
41int err = numpy::signal_from_buffer(buffer, EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE,
42&signal);
43if (err != 0) {
44ei_printf("signal_from_buffer failed (%d)\n", err);
45return;
46}
47// Run classifier
48ei_impulse_result_t result = {0};
49EI_IMPULSE_ERROR res = run_classifier(&signal, &result, debug_nn);
50if (res != EI_IMPULSE_OK) {
51ei_printf("run_classifier failed (%d)\n", res);
52return;
53}
54// Print predictions
55for (size_t ix = 0; ix < EI_CLASSIFIER_LABEL_COUNT; ix++) {
56ei_printf("%s: %.3f ", result.classification[ix].label, result.classification[ix].value);
57}
58#if EI_CLASSIFIER_HAS_ANOMALY == 1
59ei_printf("anomaly: %.3f", result.anomaly);
60#endif
61ei_printf("\n");
62}

#include <your_project_inference.h> (the exact name is shown in the Arduino
examples).
#include <Wire.h>
#include <Adafruit_ADXL345_Unified.h>
#include <your_project_inference.h> // <-- replace with your EI library header
Adafruit_ADXL345_Unified accel = Adafruit_ADXL345_Unified(12345);
// EI expects data every EI_CLASSIFIER_INTERVAL_MS
static bool debug_nn = false;
void setup() {
Serial.begin(115200);
while (!Serial) {}
Wire.begin();
delay(50);
if (!accel.begin()) {
Serial.println("ERROR: ADXL345 not detected");
while (1) { delay(10); }
}
accel.setRange(ADXL345_RANGE_4_G);
if (EI_CLASSIFIER_RAW_SAMPLES_PER_FRAME != 3) {
ei_printf("ERR: Model expects 3 axes\n");
while (1) { delay(10); }
}
Serial.println("Ready for inference (move the board: up/down/circle).");
}
void loop() {
// Capture one frame of data (features window)
float buffer[EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE] = {0};
for (size_t ix = 0; ix < EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE; ix += 3) {
uint64_t next_tick = micros() + (EI_CLASSIFIER_INTERVAL_MS * 1000);
sensors_event_t e;
accel.getEvent(&e);
buffer[ix + 0] = e.acceleration.x; // m/s^2
buffer[ix + 1] = e.acceleration.y;
buffer[ix + 2] = e.acceleration.z;
// keep precise timing
int32_t wait = (int32_t)(next_tick - micros());
if (wait > 0) delayMicroseconds(wait);
}
// Create EI signal
signal_t signal;
int err = numpy::signal_from_buffer(buffer, EI_CLASSIFIER_DSP_INPUT_FRAME_SIZE,
&signal);
if (err != 0) {
ei_printf("signal_from_buffer failed (%d)\n", err);
return;
}
// Run classifier
ei_impulse_result_t result = {0};
EI_IMPULSE_ERROR res = run_classifier(&signal, &result, debug_nn);
if (res != EI_IMPULSE_OK) {
ei_printf("run_classifier failed (%d)\n", res);
return;
}
// Print predictions
for (size_t ix = 0; ix < EI_CLASSIFIER_LABEL_COUNT; ix++) {
ei_printf("%s: %.3f ", result.classification[ix].label, result.classification[ix].value);
}
#if EI_CLASSIFIER_HAS_ANOMALY == 1
ei_printf("anomaly: %.3f", result.anomaly);
#endif
ei_printf("\n");
}

Downloadable files

Circute diagram

Connections

CKT.jpg

Documentation

Step-By- Step Tutorial PDF

Step-By- Step Tutorial PDF to refer

Step by step Tutorial for AIML Gesture Recognition with SmartElex Bharat AI Innovators Kit Powered by Arduino UNO EK R4 Wi-Fi.pdf

Comments

Only logged in users can leave comments

robuinlabs

0 Followers

•

0 Projects

Work attribution

Intro