Thursday, May 15, 2025

IOT-BASED INTELLIGENT VEHICLE ACCIDENT DETECTION AND ALERT SYSTEM

 Abstract: This paper aims to improve road safety by detecting vehicle accidents in real time and sending immediate alerts. It uses the STM32F446RE microcontroller to collect data from the MPU6050 sensor, which tracks motion and orientation, and the NEO-6M GPS module, which provides the vehicle’s exact location. When a sudden impact or unusual movement is detected, the system sounds a buzzer, stops the vehicle’s engine using a relay, and sends important details such as impact level, time, and location to the RuggedBoard A5D2X. The A5D2X then forwards this information to the ThingsBoard cloud using the MQTT protocol, enabling emergency services to be notified quickly. This project integrates sensors, microcontrollers, and cloud technology to reduce response time and improve accident location accuracy. It provides a reliable and scalable solution for intelligent transportation, with future scope for predictive analysis and GSM/GPRS-based connectivity.

Keywords: Accident Detection System, IoT, STM32F446RE, MPU6050, NEO-6M GPS, RuggedBoard A5D2X, Real-Time Monitoring, ThingsBoard, MQTT Protocol, Emergency Alert System, Cloud Connectivity, Intelligent Transportation.

 

I. INTRODUCTION

Road accidents continue to be a major global concern, causing a significant number of fatalities and severe injuries each year. Delayed detection and response times often exacerbate the impact of these accidents, resulting in increased casualties. Traditional accident reporting methods, such as relying on eyewitnesses or manual alerts, can be inefficient and unreliable in critical situations. The need for an automated, real-time accident detection system has never been more pressing.

 

To address this, this paper presents an IoT-Based Intelligent Vehicle Accident Detection and Alert System designed to enhance road safety. By leveraging the capabilities of modern embedded systems, sensors, and cloud computing, the system aims to reduce response time and improve the efficiency of emergency operations.

 

The system integrates the STM32F446RE microcontroller with an MPU6050 accelerometer and gyroscope for motion tracking, and a NEO-6M GPS module for real-time location detection.

 

When an accident occurs, the system automatically activates a buzzer for immediate audio notification, cuts off the vehicle's engine to prevent further movement, and sends accident data such as impact intensity, timestamp, and GPS coordinates via UART to the RuggedBoard A5D2X. The board processes the data and transmits it to the ThingsBoard cloud platform using the MQTT protocol for real-time monitoring and emergency notifications.

 

This integrated approach not only improves emergency response times but also ensures accurate accident location tracking. The system represents a significant step toward smarter transportation safety, with future potential for machine learning-based predictive analysis and extended communication using GSM/GPRS technology.

 

II. LITERATURE SURVEY

 

This literature review explores previous research and advancements in IoT-based vehicle accident detection systems. The focus is on sensor technologies, communication protocols, GPS-based tracking, and emergency alert mechanisms. These studies help in identifying technological gaps and in shaping the proposed system design.

 

1. Accident Detection and Reporting Using GPS, GPRS, and GSM (2012, IEEE)

This study utilizes GPS to monitor vehicle speed and detect accidents based on sudden speed drops. When a significant deceleration is identified, the system assumes a potential crash and sends the GPS location and time via GSM to an alert service center. A 5-second window is provided to cancel false alarms. This system eliminates the need for speedometers by directly calculating speed from GPS data.

 

2. Real-Time Detection and Reporting of Vehicle Collisions (2017, IEEE)

This work uses an accelerometer and gyroscope to detect orientation changes and impacts. When a threshold angle or acceleration force is exceeded, the system identifies a collision or vehicle rollover. Alerts are then sent via GSM to family members and the nearest hospital. GPS data is used to confirm location accuracy and eliminate false positives.


3. Vehicle Accident Detection Using GSM, GPS, and Sensors (2019, IRJET)

This paper employs a piezoelectric sensor to detect collisions based on voltage spikes generated during impact. When the voltage exceeds a set threshold, the system captures GPS coordinates and transmits them via GSM to the rescue team. A Google Maps module is used to visualize accident locations. An emergency switch is included to prevent false alerts.

 

Table.1. Comparison Between Research Papers

 

III. METHODOLGY

The intelligent vehicle accident detection and alert system is designed using STM32F446RE as the main controller along with MPU6050 sensor, NEO-6M GPS module, DC motor, buzzer, 16x2 I2C LCD, RuggedBoard A5D2X, and Ethernet cable. The system is explained using a block diagram. The STM32 board is the main focus for sensor data and control. For accident detection, we used the MPU6050 sensor which detects acceleration and angular motion in different directions like front, back, left, right, and upside-down. Threshold values are set to identify accidents from regular motion. The NEO-6M GPS module gets the location of the vehicle and sends it when an accident is detected. The DC motor is connected and used to simulate vehicle movement. A buzzer is connected to STM32 pins PA5 and PA6 to give sound alerts during an accident. The LCD shows messages like "Vehicle Running - No Accident" and "Vehicle Stopped - Accident Detected". The data is sent from STM32 to RuggedBoard A5D2X through UART. The RuggedBoard sends accident alerts and GPS data to the cloud using MQTT over Ethernet. This system helps in real-time accident detection and cloud-based alerting.Power is supplied to the components through a regulated source to ensure stable operation.The firmware includes logic to continuously compare motion data against set thresholds.This integration of hardware and software ensures a compact, responsive, and intelligent embedded system.

 

Figure.1. Block Diagram of Vehicle accident detection and alert system

 

IV. MODELLING OF THE SYSTEM

Firmware is a computer program embedded into hardware and plays a vital role in controlling the behavior of embedded systems. In our project, the embedded firmware is developed for the STM32F446RE microcontroller, which is responsible for reading sensor data, detecting accidents, and controlling other components. This firmware acts as the bridge between hardware and software, similar to how an operating system functions in a computer. It runs continuously to monitor data from connected modules and make real-time decisions based on pre-defined conditions.

 

The STM32 executes the logic that controls the MPU6050 sensor, which detects abnormal motion such as sudden impact, tilting, or flipping. Based on acceleration and gyroscope data, the system identifies whether an accident has occurred. Once an accident is detected, the STM32 triggers the buzzer connected to PA5 and PA6 for alerting, updates the message on the 16x2 I2C LCD, and stops the motor to simulate vehicle shutdown. At the same time, the GPS NEO-6M module fetches the location data (latitude and longitude).

 

This information is transmitted through UART to the RuggedBoard A5D2X, which sends the accident data to the ThingsBoard cloud via MQTT using Ethernet. This enables remote monitoring and real-time alerts for emergency response. The overall flow of the system is continuous: STM32 reads data from the MPU6050 in a loop. If no accident is detected, the vehicle continues to run normally. If an accident is detected, the motor is stopped, alerts are triggered, and location data is sent to the cloud.To ensure robustness, error-checking mechanisms are included in the UART communication.
The firmware is optimized for low-latency response, ensuring immediate reaction during critical events.
Modular code design also allows easy firmware updates and scalability for future enhancements.

 

 

Figure.2. Flowchart of the overall system.

 

 

 V. DATA COLLECTION

Efficient data collection and processing are essential for accurate decision-making and enhanced system performance in vehicle safety applications. In the proposed Intelligent Automatic Vehicle Accident Detection System, sensor data is collected using the STM32F446RE microcontroller, which interfaces with critical components such as the MPU6050 accelerometer/gyroscope, NEO-6M GPS module, relay module, and buzzer.

The system employs a polling technique, where the microcontroller continuously reads sensor values to monitor sudden changes in movement or orientation. Although effective, this method could be optimized in future versions by incorporating interrupt-based techniques, which would improve responsiveness and reduce power consumption by reacting only to significant events.

This paper addresses a significant real-world issue: reducing fatalities and injuries from road accidents due to delayed emergency responses. According to the World Health Organization (WHO), millions of lives are lost each year in traffic-related incidents. By enabling real-time accident detection and transmitting alerts with precise location data, the system supports the United Nations Sustainable Development Goal (UN SDG) 3.6, which aims to halve the number of global road traffic deaths and injuries.

When the MPU6050 detects unusual movement or an impact exceeding a defined threshold, the STM32F446RE immediately activates a buzzer, cuts off the engine via a relay, and transmits critical data including acceleration, gyroscope readings, and GPS coordinates via UART communication to the RuggedBoard A5D2X. The RuggedBoard then forwards this information to the ThingsBoard IoT platform.

In future enhancements, the system could integrate machine learning algorithms to predict risky driving behavior and biometric sensors to monitor the driver’s physical condition after an accident. This intelligent and connected architecture highlights the transformative potential of embedded IoT in building smarter, safer transportation ecosystems.

 

Table.2. Data log

 

 

VI. PROPOSED SYSTEM

The proposed system is an IoT-enabled, real-time Intelligent Vehicle Accident Detection and Alert System designed to automatically detect vehicle collisions and promptly notify emergency services and relevant stakeholders. This system aims to minimize the delay between the occurrence of an accident and the emergency response, thereby reducing fatalities and the severity of injuries. At the core of the system is the STM32F446RE microcontroller, which manages data acquisition, processing, and control logic. The system interfaces with the following key components:

 

a.      STM32F446RE

The STM32F446RE microcontroller serves as the central processing unit of the system, handling critical tasks such as data acquisition, processing, and communication. It continuously monitors data from the MPU6050 sensor and the NEO-6M GPS module. When sensor data indicates a potential accident such as a sudden impact, extreme tilt, or abrupt movement exceeding predefined threshold values, the STM32F446RE initiates the accident response mechanism. This includes activating the buzzer, shutting down the engine via the relay, and transmitting crucial data, such as location and movement details, to the RuggedBoard A5D2X for further processing and cloud integration.

The STM32F446RE features high-performance ARM Cortex-M4 architecture, enabling fast and efficient signal processing. It supports multiple UART interfaces, allowing seamless communication with peripherals like GPS and other external boards. Its low-power consumption and integrated timers make it suitable for real-time embedded applications in automotive safety systems.

b.      MPU6050 SENSOR

The MPU6050 sensor is essential for the accident detection mechanism of the system. This 6-axis motion sensor integrates a 3-axis accelerometer and a 3-axis gyroscope, enabling it to continuously monitor the vehicle’s motion and orientation. In the event of a collision, rapid deceleration, or abnormal tilt, the sensor sends data to the STM32F446RE microcontroller. If the data surpasses predefined thresholds, the system interprets this as an accident. Upon detection, the system triggers immediate actions, including activating the buzzer, shutting down the engine, and sending critical data to the cloud for further analysis. The sensor’s precision ensures reliable real-time detection of various accident scenarios, including frontal collisions, side impacts, and rollovers.

 

c.      RUGGED BOARD A5D2X

The RuggedBoard A5D2X acts as an intermediary between the STM32F446RE microcontroller and the cloud platform. It receives accident-related data such as acceleration, orientation, and GPS coordinates from the microcontroller via UART communication. The RuggedBoard subsequently transmits this data to a cloud-based platform, such as ThingsBoard, for real-time monitoring and further analysis. The cloud integration allows emergency responders to access critical information promptly, enhancing response time and coordination.

 

d.      CLOUD MONITORING

Cloud monitoring is a key feature of the system, enabling remote access to both real-time and historical accident data. After the STM32F446RE microcontroller processes data from the MPU6050 sensor and the NEO-6M GPS module, the data is transmitted to the RuggedBoard A5D2X for forwarding to a cloud platform. Cloud platforms, such as ThingsBoard, allow for secure storage, visualization, and analysis of the data, providing emergency services with immediate access to important accident information, including vehicle status, location, and impact details. This cloud-based solution ensures efficient monitoring, faster response times, and seamless coordination with emergency teams.

 

e.      ALERTING AND CONTROL MECHANISMS

In the event of an accident, the system activates a series of alerting and control mechanisms. A buzzer emits an audible alert to notify passengers and bystanders of the detected collision, while a relay module, in conjunction with a DC motor, is used to shut down the vehicle’s engine and prevent further damage or hazards. Simultaneously, a 16x2 I2C LCD display provides real-time status updates, displaying messages such as "Vehicle Running - No Accident" or "Vehicle Stopped - Accident Detected." These mechanisms ensure that both vehicle occupants and nearby individuals are immediately alerted, and the vehicle is safely brought to a stop.


Figure.3. Connection diagram of Accident detection and Alert system


 

VII. IMPLEMENTATION

a.      Sensor and GPS Module Connections

The MPU6050 accelerometer and gyroscope sensor is connected to the STM32 via I2C (PB6 and PB7) to detect movements, vibrations, or crashes. The same I2C lines are shared with the 16x2 LCD display, which shows system messages. The NEO-6M GPS module connects to the STM32 via UART (PA3 for RX and PA2 for TX) to provide real-time location data, including latitude and longitude, which are crucial for determining the accident location.

 

b.      Relay, DC Motor, and Buzzer Connections

The relay module, connected to STM32 pin PA6, controls the DC motor, simulating the vehicle engine. In case of an accident, the STM32 deactivates the relay to stop the motor. The buzzer, connected to PA5, sounds an alert during an accident.

 

c.      UART Communication with RuggedBoard

The STM32 communicates with the RuggedBoard A5D2X via UART. The TX (PA9) and RX (PA10) pins of the STM32 send sensor data and GPS coordinates (latitude and longitude) to the RuggedBoard, which then sends the data to the ThingsBoard Cloud using MQTT over Ethernet.

 

d.      Working of the System

The STM32 monitors the MPU6050 sensor for any abnormal motion. If an accident occurs, the system stops the motor (PA6), activates the buzzer (PA5), and shows "Accident Detected – Vehicle Stopped" on the LCD. The system also sends the GPS location (latitude and longitude) to the RuggedBoard via UART. If no accident occurs, the vehicle runs normally, and the LCD shows "Vehicle Running – No Accident."

Figure.4.Interfacing controller with all other module

 

VIII. RESULTS AND DISCUSSION

This chapter presents the methodology adopted for the design, development, and implementation of the IoT-Based Intelligent Automatic Vehicle Accident Detection System. It details the research design, data collection techniques, system development, testing procedures, and performance evaluation criteria to ensure system accuracy and reliability. The study follows an experimental research approach, where a prototype was developed and tested under real-world conditions.

1. Reading Data from MPU-6050 Module

The MPU-6050 is a combined 3-axis gyroscope and 3-axis accelerometer based on MEMS (Micro-Electro-Mechanical Systems) technology. It is connected to the STM32F446RE microcontroller using the I2C communication protocol. To verify the connection, I2C communication was tested by writing code that initializes the I2C and UART, and reads data from the relevant registers.

 

Figure.5. MPU 6050 Data reading

 

2. Validating UART Communication Between STM32 and RuggedBoard   

UART (Universal Asynchronous Receiver/Transmitter) communication is established between the STM32F446RE microcontroller and the RuggedBoard A5D2X for transferring sensor and GPS data. The STM32 uses two UART channels: UART1 for GPS data and UART2 for communication with the RuggedBoard.The data sent includes gyroscope and accelerometer values, as well as GPS coordinates.This real-time data is processed and displayed on the RuggedBoard, which further enables cloud connectivity.The use of dual UART channels ensures efficient and parallel handling of both sensor and location data streams.

Figure.6. UART Communication

 

3. Cloud Data Logging and Visualization

Cloud data logging and visualization play a crucial role in monitoring the IoT-Based Intelligent Vehicle Accident Detection System. The system transmits real-time sensor data, including vehicle status and location, to the ThingsBoard Cloud for remote monitoring and analysis.

 

Figure.7. Cloud Data logging

 

4. Location Data Reading from GPS Module

The U-blox NEO-6M GPS module is used to acquire real-time location data. It communicates with the STM32F446RE microcontroller over UART and sends data in NMEA sentence format, which includes latitude, longitude, and time. The onboard LED provides visual feedback, blinking rapidly when searching for satellites and once per second when a position fix is acquired. The STM32 parses this data, extracts the GPS coordinates, and transmits the information to the RuggedBoard A5D2X for further processing and remote visualization through the ThingsBoard Cloud.

 

Figure.8. GPS Data reading.

5. Verifying LCD Display Output

The system includes a 16x2 LCD connected via an I2C module, which minimizes pin usage. The LCD displays system messages based on real-time sensor input. During normal operation, it shows "Vehicle Running – No Accident." In case of an accident, the LCD displays "Accident Detected – Vehicle Stopped."

 

Figure.7. LCD Display Output Testing

6. Full System Integration Testing

After individual module testing, the entire system was integrated and tested for operation. This included the MPU-6050 sensor, GPS module, buzzer, relay, LCD display, and UART communication interface. When an accident was simulated by shaking or tilting the sensor, the system responded by activating the buzzer, stopping the motor through the relay, updating the LCD display, and sending all sensor and location data to the RuggedBoard for remote monitoring.

 

Figure.9. Full System Integration Test

 

7. Scalability and Future Enhancement

For future enhancements, improvements in communication redundancy, sensor capabilities, user interaction, and data management will be explored. These enhancements aim to increase the performance, reliability, and scalability of the IoT-Based Intelligent Vehicle Accident Detection and Alert System.
Integration with cloud-based analytics platforms can enable predictive insights and accident pattern recognition. Incorporating machine learning algorithms may enhance accuracy in accident detection and false-positive filtering. Additionally, support for mobile notifications and real-time alerts can further improve user responsiveness and emergency handling.

IX. CONCLUSION

The IoT-Based Intelligent Automatic Vehicle Accident Detection System offers an effective solution to improve road safety by detecting accidents in real-time and sending alerts quickly. This system combines key components like the STM32F446RE microcontroller, MPU6050 motion sensor, and NEO-6M GPS module to accurately detect accidents and share location details with emergency services.The MPU6050 sensor allows the system to detect accidents from multiple directions front, back, sides, and even rollovers. The relay module automatically stops the engine during an accident, helping prevent further damage.

REFERENCE

[1]    Vijayan N., IoT-based intelligent accident detection and emergency alert system using Bluetooth and LoRa communication, 2025 International Conference on Smart Technologies and Systems (ICSTS).

[2]    Amit Kumar, IoT based automatic vehicle accident detection and rescue system, HashStudioz Technologies, Technical Report, April 2024.

[3]    Shafin Talukder; Mohammed Imran; Kazi A. Rahman, Vehicle collision detection & prevention using VANET based IoT with V2V, May 2022.

[4]    Victor Adewopo; Nelly Elsayed, Smart city transportation: Deep learning ensemble approach for traffic accident detection , October 2023.

[5]    S. Shruthi; M. Pooja; S. Akash, Accident detection and alerting system using Arduino and GPS module, 2021 International Conference on Smart Electronics and Communication (ICOSEC).

[6]    M. Hossain; M. Moniruzzaman; T. Rahman, Real-time road accident detection using IoT and cloud services, 2020 International Conference on Electrical, Computer and Communication Engineering (ECCE).

[7]    Syed S. Rizvi; Waqas Anwar; Usama Sadiq, Design and implementation of an IoT-based accident detection and notification system using GSM and GPS, 2020 3rd International Conference on Computing, Mathematics and Engineering Technologies (iCoMET).

[8]    T. A. Hossain; M. H. Hossain; S. Roy, Smart vehicle accident detection and tracking system using GPS and GSM module, 2021 IEEE Global Conference for Advancement in Technology (GCAT).

[9]    R. Gopal; M. Kumar; V. Gupta, IoT based accident detection system using NodeMCU, 2022 6th International Conference on Trends in Electronics and Informatics (ICOEI).

[10] M. Rahman; A. Chowdhury; N. Hossain, Vehicle accident detection and alert system using accelerometer and GPS module, International Journal of Computer Applications, vol. 182, no. 28, pp. 1-5, January 2019.

1 comment:

IOT-BASED INTELLIGENT VEHICLE ACCIDENT DETECTION AND ALERT SYSTEM

  Abstract: This paper aims to improve road safety by detecting vehicle accidents in real time and sending immediate alerts. It uses the STM...