1. Introduction

Natural disasters, particularly earthquakes, pose significant threats to human life and infrastructure worldwide. According to the United Nations Office for Disaster Risk Reduction (UNDRR), earthquakes caused approximately 750,000 deaths globally between 1998 and 2017, with economic losses exceeding $600 billion [1]. In India, the seismic hazard map classifies approximately 59% of the land area as susceptible to moderate to severe earthquakes, with regions such as the

Himalayan belt, Northeast India, and parts of Gujarat and Bihar falling under high-risk zones [2].

Traditional seismic monitoring networks, such as those operated by the Indian Meteorological Department (IMD) and the United States Geological Survey (USGS), employ highly sensitive seismometers and sophisticated data acquisition systems. While these systems provide accurate and reliable data, they are prohibitively expensive for small-scale deployment, costing anywhere from ₹5 lakhs to ₹50 lakhs per station [3]. Consequently, there exists a significant gap in affordable, accessible vibration detection solutions for educational institutions, small laboratories, and residential safety applications.

1.1 Problem Statement

The primary challenges in developing low-cost vibration detection systems include:

1.       High-precision seismometers are expensive, while low-cost accelerometers may lack sufficient sensitivity for detecting subtle ground motions.

2.       Effective early warning systems require minimal latency between vibration detection and alert generation.

3.       Distinguishing between hazardous vibrations and everyday disturbances (e.g., passing vehicles, footsteps) remains challenging.

4.       Continuous monitoring systems must operate efficiently, particularly for battery-powered or remote installations.

5.       Presenting vibration data in an accessible format for non-expert users is a key challenge.

1.2 Objectives

The primary objectives of this work are:

1.         To design a low-cost vibration detection system using readily available components

2.         To implement real-time monitoring of vibration intensity along three orthogonal axes

3.         To develop threshold-based alert generation with both visual (LCD) and audible (buzzer) outputs

4.         To evaluate system performance under controlled vibration conditions

5.         To demonstrate the system's applicability for educational and safety monitoring applications

2. Literature review

Vibration detection and seismic monitoring have been active areas of research for several decades. Table 1 summarizes key contributions from existing literature relevant to low-cost implementation approaches.

The present work differentiates itself by providing a comprehensive, reproducible implementation with quantitative performance evaluation, threshold-based alerting, and integrated visual display specifically designed for educational accessibility and residential safety applications.

Table 1. Summary of existing work

3. System architecture and components

3.1 Overall System Design

The proposed vibration detection system employs a modular architecture comprising four functional units: (1) sensing unit, (2) processing unit, (3) display unit, and (4) alert unit. Figure 1 presents the block diagram of the system.

Fig. 1. Block Diagram of Vibration Detection System

The sensing unit utilizes the ADXL335 triple-axis accelerometer to measure vibration intensity along the X, Y, and Z axes. The analog voltage outputs from the sensor are fed to the analog input pins of the Arduino UNO microcontroller, which serves as the processing unit. The microcontroller continuously samples these analog values, converts them to digital readings using its built-in 10-bit analog-to-digital converter (ADC), and compares them against user-defined threshold values. When vibration intensity exceeds the threshold on any axis, the processing unit triggers the alert unit (piezoelectric buzzer) and updates the display unit (16×2 I2C LCD) with the current vibration status and intensity values.

3.2 Component Specifications and Selection

3.2.1 ADXL335 Triple-Axis Accelerometer

The ADXL335 is a small, low-power, complete 3-axis accelerometer with analog voltage outputs. Manufactured by Analog Devices, it measures acceleration with a minimum full-scale range of ±3g and provides a sensitivity of approximately 300 mV/g at 3V supply voltage [11]. Key specifications include:

·       Operating Voltage: 1.8V to 3.6V DC (3.3V recommended for optimal accuracy)

·       Supply Current: 350 µA (typical)

·       Measurement Range: ±3g (minimum)

·       Sensitivity: 270–330 mV/g (300 mV/g typical at 3V)

·       Bandwidth: 0.5 Hz to 1600 Hz (user-selectable via external capacitors)

·       Output Type: Analog voltage for each axis (X, Y, Z)

·       Operating Temperature: -40°C to +85°C

The ADXL335 operates on the principle of micro-electromechanical systems (MEMS) technology. Inside the device, a small proof mass is suspended by polysilicon springs. Acceleration causes the proof mass to deflect, changing the differential capacitance between fixed plates and plates attached to the mass. A CMOS signal conditioning circuit converts this capacitance change to an analog output voltage proportional to acceleration [12].

The ADXL335 was selected for this application due to its low cost (₹350), low power consumption (suitable for battery operation), sufficient sensitivity for vibration detection, ease of interfacing (analog outputs), and wide availability.

Fig. 2. Pinout Diagram of ADXL335 Sensor

ADXL335 Pin Connections:

·           VCC → 3.3V supply (from Arduino 3.3V pin)

·           GND → Ground (Arduino GND)

·           X OUT → Analog pin A0 (Arduino)

·           Y OUT → Analog pin A1 (Arduino)

·           Z OUT → Analog pin A2 (Arduino)

3.2.2 Arduino UNO Microcontroller

The Arduino UNO serves as the central processing unit of the proposed system. It is based on the ATmega328P microcontroller, which operates at a clock frequency of 16 MHz [13]. Key specifications include:

·           Digital I/O Pins: 14 (6 with PWM capability)

·           Analog Input Pins: 6 (10-bit resolution, 0–5V range)

·           Flash Memory: 32 KB (0.5 KB used by bootloader)

·           SRAM: 2 KB

·           EEPROM: 1 KB

·           Operating Voltage: 5V

·           Input Voltage (recommended): 7–12V

The Arduino UNO was selected for its ease of programming, extensive library support, large community documentation, built-in 10-bit ADC for sensor interfacing, and proven reliability in educational and prototyping applications [14].

3.2.3 16 × 2 I2C LCD Display

The 16×2 I2C LCD display is a character display module capable of showing 16 characters per line across 2 lines. It incorporates an I2C (Inter-Integrated Circuit) backpack based on the PCF8574 I/O expander, which converts I2C serial communication to parallel commands required by the LCD's HD44780 controller [15]. Key specifications include:

·       Display Format: 16 characters × 2 lines

·       Interface: I2C (SDA, SCL) – only 2 wires required

·       Operating Voltage: 5V DC

·       Backlight: LED (adjustable contrast via potentiometer)

·       I2C Address: Typically 0x27 or 0x3F (user-configurable)

·       The I2C interface significantly reduces wiring complexity compared to standard parallel LCDs (which require 6–10 connections), improving reliability and simplifying assembly.

3.2.4 Piezoelectric Buzzer

A piezoelectric buzzer is employed as the audible alert device. When a voltage is applied across the piezoelectric element, it deforms mechanically, generating sound waves [16]. Key specifications include:

·           Operating Voltage: 3–12V DC (5V compatible)

·           Resonant Frequency: 2–4 kHz

·           Sound Output: 85–95 dB at 10 cm

·           Current Consumption: <30 mA at 5V

The buzzer is directly driven by a digital output pin of the Arduino UNO, with a transistor driver circuit for improved current handling if required.

3.2.5 Breadboard

A solderless breadboard is used for prototyping the circuit before permanent assembly. The breadboard consists of interconnected holes arranged in rows and columns: horizontal power rails (typically marked red for positive and blue for ground) and vertical terminal strips for component connections [17].

3.3 Cost Analysis

Table II presents the detailed cost breakdown of the prototype components. The total cost of ₹900 (approximately USD 11) demonstrates the economic viability of the proposed system, making it accessible for educational institutions and residential applications.

Table 2. Cost and specifications

4. Hardware implementation and software algorithm

4.1 Circuit Schematic and Connections

The complete circuit connections are as follows:

ADXL335 to Arduino UNO:

·       ADXL335 VCC → Arduino 3.3V (not 5V, as ADXL335 operates at 3.3V)

·       ADXL335 GND → Arduino GND

·       ADXL335 X OUT → Arduino A0

·       ADXL335 Y OUT → Arduino A1

·       ADXL335 Z OUT → Arduino A2

I2C LCD to Arduino UNO:

·                  LCD VCC → Arduino 5V

·                  LCD GND → Arduino GND

·                  LCD SDA → Arduino A4 (SDA)

·                  LCD SCL → Arduino A5 (SCL)

Buzzer to Arduino UNO:

·                  Buzzer Positive (+) → Arduino Pin 8

·                  Buzzer Negative (-) → Arduino GND

Optional: LCD Contrast Adjustment

·                  Potentiometer middle pin → LCD V0 pin

·                  Potentiometer outer pins → 5V and GND

Fig. 3. Complete Hardware Model

Figure 3 shows the assembled prototype with the ADXL335 sensor, Arduino UNO, I2C LCD display, buzzer, and breadboard connections.

4.2 System Operation Principle

The operational logic of the system is as follows:

1. Initialization Phase: Upon power-up, the Arduino UNO initializes the I2C LCD display, displays a welcome message, and calibrates the baseline vibration levels by averaging 100 readings from each axis.

2. Continuous Monitoring Phase: The microcontroller continuously reads the analog voltage values from the X, Y, and Z outputs of the ADXL335 sensor. Each reading is converted to a digital value (0–1023) corresponding to 0–5V input. These raw values are then converted to acceleration in g-units using the sensor's sensitivity (approximately 330 mV/g at 5V operation).

3. Threshold Comparison: The processed acceleration values are compared against user-defined threshold values (typically set at 0.5g above baseline for safety applications).

4. Alert Generation: If any axis reading exceeds the threshold, the system:

·         Activates the buzzer (digital write HIGH on Pin 8)

·         Updates the LCD to display "VIBRATION DETECTED" along with the current acceleration values

·         Maintains the alert state for a configurable duration (typically 3 seconds)

5. Reset and Recovery: When vibration intensity returns below the threshold for a sustained period (e.g., 2 seconds), the system deactivates the buzzer and resumes normal monitoring display.

4.3 Software Algorithm

The Arduino reads analog values from the ADXL335 sensor (X, Y, Z axes) and compares them against predefined thresholds. If any axis exceeds the threshold, the buzzer activates and the LCD displays "VIBRATION DETECTED"; otherwise, the system shows "Status: Normal." The system performs automatic baseline calibration at startup to ensure accurate detection.

4.4 Calibration Procedure

Proper calibration is essential for reliable operation. The following calibration steps are performed:

1. Baseline Acquisition: With the sensor stationary on a level surface, the system acquires 100 samples from each axis and computes the average. These baseline values represent the static gravitational component (approximately 1g on the Z-axis when horizontal).

2. Threshold Setting: The threshold values are set experimentally based on:

o     Ambient vibration levels in the deployment environment

o     Desired sensitivity (lower threshold = more sensitive, higher false alarm rate)

o     Application requirements (e.g., structural monitoring vs. intrusion detection) For seismic monitoring applications, a threshold of 0.2–0.5g deviation from baseline is recommended.

5. Results and discussion

5.1 Experimental Setup

The prototype was tested under controlled laboratory conditions to evaluate its performance across various vibration scenarios:

·       Test Case 1: No vibration (baseline stability test)

·       Test Case 2: Light tapping on the mounting surface (simulating footsteps)

·       Test Case 3: Moderate impact (simulating door slam or falling object)

·       Test Case 4: Continuous vibration (simulating machinery operation)

·       Test Case 5: Simulated seismic activity (shaking the sensor platform)

Each test was repeated 20 times, and the system's response (detection success, false alarm, response time) was recorded.

5.2 Functional Performance

Table 3 summarizes the experimental results.

Table 3. Experimental results

The system achieved an overall detection rate of 96.25% across all test cases, with an average response time of 70 milliseconds. False alarms occurred in 4% of trials, primarily due to:

1.         Electromagnetic interference from nearby electronic devices (2 instances)

2.         Accidental contact with the breadboard during testing (2 instances)

5.3 Sensitivity Analysis

The sensitivity of the ADXL335 sensor was characterized by applying known accelerations (using a calibrated shaker table) and measuring the output voltage. The results are presented in Table IV.

Table 4. Sensitivity characterization

The measured sensitivity of 300 mV/g matches the manufacturer's specification, confirming proper sensor operation.

5.4 Comparison with Existing Work

Table 5 compares the proposed system with existing low-cost vibration detection implementations.

Table 5. Comparative analysis

The proposed system offers comparable or improved functionality relative to existing low-cost implementations, with the added benefit of I2C-based display (reducing wiring complexity) and automated baseline calibration.

5.5 Limitations

The following limitations were identified during the study:

1.       ADXL335's bandwidth (up to 1,600 Hz) may miss very high-frequency vibrations despite being sufficient for seismic monitoring (0.5–20 Hz).

2.       Current prototype lacks non-volatile storage for recording vibration events over time.

3.       System lacks wireless connectivity, requiring physical presence for alert observation.

4.       Fixed threshold does not adapt to changing environmental conditions like diurnal ambient vibration variations.

5.       Prototype requires continuous USB/DC power as battery operation with power management was not implemented.

6. Conclusion and future scope

6.1 Conclusion

This paper successfully presented the design, implementation, and evaluation of a low-cost vibration detection system for seismic monitoring and safety applications. The system integrates an ADXL335 triple-axis accelerometer for vibration sensing, an Arduino UNO microcontroller for data acquisition and processing, a 16×2 I2C LCD for real-time display of vibration intensity, and a piezoelectric buzzer for audible alert generation.

6.2 Future Enhancements

Future iterations of the system can be enhanced by integrating IoT capabilities (ESP8266/ESP32) and SD card logging for wireless remote monitoring, cloud storage, and long-term vibration data analysis. An intelligent multi-tier alert system combined with machine learning would enable classification of vibration sources (earthquake vs. construction vs. footsteps), reducing false alarms while providing nuanced risk assessment. Additionally, battery management with sleep functionality and Bluetooth smartphone notifications would enable flexible remote deployment, while networking multiple sensors would allow vibration source triangulation for larger facilities.

6.3 Applications

The proposed system is well-suited for educational laboratories to teach sensor interfacing and embedded systems, residential safety for early warning of structural vibrations and unauthorized entry, industrial vibration monitoring for rotating machinery, small-scale seismic monitoring as a low-cost supplement to professional seismometers, structural health assessment for small buildings and bridges, and security applications such as detecting unauthorized handling of museum exhibits.

Acknowledgment

The authors sincerely thank Prof. (Dr.) Pankaj Jha, Head of ECE Department, IIMT College of Engineering, Greater Noida, for his invaluable guidance and support. The authors also acknowledge the faculty members and technical staff for providing laboratory facilities and technical assistance. Finally, thanks to their B.Tech. batchmates for their constructive feedback and collaboration.