INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
Emotion BasedAutonomous Driving Control Using Multi Sensor  
Integration for Enhanced EV Experience  
B. Noorul Hamiitha¹ M.Girija ² K. Durga Devi3 P.Pavithra4 K. Ajith5  
¹
,2,3,4Fatima Michael College of Engineering and Technology  
DOI: https://doi.org/10.51583/IJLTEMAS.2026.150500161  
Received: 30 May 2026; Accepted: 06 June 2026; Published: 10 June 2026  
ABSTRACT  
This paper presents an emotion-based autonomous driving control system that integrates real-time physiological  
monitoring with autonomous vehicle technology to enhance driving safety and personalization. The system  
continuously evaluates the driver’s emotional state using sensors that measure pulse rate, oxygen saturation, and  
body temperature, dynamically adjusting the vehicle’s driving mode between manual and autonomous according  
to the detected condition. By autonomously assuming control during stress or fatigue, the system mitigates the  
risk of human error, particularly in high-stress scenarios, while providing real-time feedback to maintain driver  
trust. Experimental results demonstrate high accuracy in emotional state classification and reliable autonomous  
driving performance. However, external factors affecting sensor readings and minor delays in mode transitions  
at high speeds highlight areas for further optimization. Future improvements, including enhanced sensor  
precision, faster mode-switching algorithms, and a more robust classification model, could further increase  
system effectiveness. This emotion-aware approach represents a significant advancement in human-centered  
autonomous driving, offering safer, adaptive, and more comfortable driving experiences.  
Keywords: Driver fatigue detection, Stress-aware vehicle systems, Sensor fusion algorithms, Adaptive human–  
machine interaction, Real-time mode switching  
INTRODUCTION  
The advent of Electric Vehicles (EVs) and autonomous driving technologies has significantly reshaped the  
landscape of modern transportation. As EVs continue to gain popularity due to their environmental benefits and  
sustainability, the integration of autonomous driving features promises to further enhance driving efficiency,  
safety, and comfort. However, while autonomous systems are adept at navigating complex environments, they  
still lack a critical component: the ability to adapt to the driver’s emotional state and well-being in real-time.  
Research as shown that emotional factors like stress, fatigue, and anxiety can significantly affect a driver’s  
performance, response time, and decision-making ability, thereby increasing the risk of accidents or errors.  
Therefore, integrating emotion recognition systems into autonomous vehicles could be a pivotal step towards  
a safer and more personalized driving experience.  
Traditional autonomous systems rely heavily on environmental data, such as road conditions, obstacles, and  
traffic signals, to make driving decisions. While this is essential for vehicle navigation, it overlooks a key aspect:  
the human driver. The emotional and physiological state of the driver plays a crucial role in their ability to  
respond to dynamic situations on the road. Stress, anxiety, fatigue, and other emotional states can impair a  
driver’s judgment, making them more susceptible to errors. Despite advancements in vehicle automation, no  
current system adequately addresses these emotional factors, which remain a significant research gap.  
To fill this gap, this research proposes an emotion-based autonomous driving control system that integrates  
real-time monitoring of the driver’s physiological signals, such as pulse rate, oxygen saturation (SpO2), and  
body temperature, to detect emotional states and adjust the vehicle’s driving mode accordingly. By  
continuously assessing the driver’s emotional condition, the system can automatically switch between manual  
Page 2024  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
and autonomous driving modes, ensuring that the vehicle adapts to the driver's needs and promotes a safer  
driving environment. This system represents a novel approach to making autonomous vehicles more human-  
centric, by not only responding to external conditions but also considering the internal emotional state of the  
driver.  
Survey  
S. Springer et al. (2025) review affective human–vehicle interaction, emphasizing how emotion recognition  
impacts driver acceptance of autonomous systems. Techniques include facial expression analysis, EEG signals,  
heart rate variability, and voice tone detection. Emotion-aware systems can adjust driving style (e.g., calming  
acceleration for stressed drivers, alertness checks for fatigued drivers).  
Qian et al. (2025, MDPI) provide a structured overview of multi-sensor fusion, integrating camera, LiDAR,  
radar, and ultrasonic sensors. Fusion strategies include BEV (Bird’s Eye View) representation and cross-  
modal attention, enabling robust perception in complex environments. Challenges include spatio-temporal  
misalignment, domain shifts, and interpretability.  
Wei et al. (2025, arXiv) categorize fusion into data-level, feature-level, and decision-level  
approaches. Highlights the role of deep learning and Vision-Language Models (VLMs) in  
enhancing adaptability. Demonstrates how fusion improves adaptive cruise control, lane keeping,  
and collision avoidance, especially under adverse weather.  
Physiological sensors (EEG, ECG, GSR) and behavioral sensors (camera, microphone) are integrated  
to detect driver states. Multi-sensor integration ensures higher accuracy compared to single-sensor  
systems. Example: Combining facial recognition + heart rate monitoring reduces false positives  
in stress detection.  
IoT platforms transmit sensor data to the cloud for real-time monitoring and predictive  
analytics.Mobile apps provide drivers with emotion-aware feedback, adjusting EV performance for  
comfort and safety.Integration with EV systems enhances range management, battery safety, and  
personalized driving experience.  
Proposed System  
The proposed system offers an advanced, emotion-based autonomous driving solution that prioritizes the well-  
being of the driver by continuously monitoring physiological indicators of stress. It integrates multiple sensors—  
each designed to track key vital parameters such as oxygen saturation (SpO2), pulse rate, and body temperature.  
These sensors are embedded into a custom Arduino-based control unit, which acts as the central hub for data  
collection, processing, and decision-making.  
The real-time data gathered by the sensors is transmitted to a Multi-Layer Perceptron (MLP) classifier, a type of  
artificial neural network that is trained to assess the driver's emotional state based on the monitored parameters.  
By analyzing the physiological data, the classifier determines the stress level of the driver—whether the  
individual is calm, stressed, or experiencing high levels of anxiety. The MLP model is specifically designed to  
interpret subtle changes in the physiological signals that might otherwise go unnoticed, such as an increased  
pulse rate or a drop in oxygen saturation, which are typical indicators of emotional distress.  
Upon receiving the classification results from the MLP, the system responds dynamically. If the driver is  
identified as being under stress or showing signs of fatigue, the system automatically transitions the vehicle to  
autonomous driving mode, thereby taking over control of the vehicle and allowing the driver to relax.  
Conversely, if the driver is in a calm state, the system can remain in manual mode, giving the driver full  
controlThe system also has a feedback loop, where it continuously monitors and adjusts its operation based on  
real-time changes in the driver’s emotional state  
Page 2025  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
Figure: Proposed system  
Sensor Module  
The Sensor Module plays a critical role in continuously monitoring the driver’s physiological signals, which  
are key indicators of emotional and stress levels.  
Arduino Control Unit  
The Arduino Control Unit is the system’s core, integrating sensors and communication. It receives sensor data,  
performs initial filtering, and transmits it for further processing. It plays a decision-making role, guided by the  
classifier module’s results. Based on these results, it sends signals to switch between manual and autonomous  
modes. This ensures seamless integration and real-time adaptation to the driver’s emotional state.  
Data Processing & Classification Module  
The Data Processing & Classification Module is central to the emotional state detection process. This module  
uses a Multi-Layer Perceptron (MLP) classifier, a machine learning model that processes the sensor data to  
evaluate the driver’s emotional condition. The classifier is trained on a dataset that includes various physiological  
signals corresponding to different emotional states—calm, stressed, or anxious  
Decision-Making Module  
The Decision-Making Module interprets the output from the MLP classifier to determine the appropriate driving  
mode for the vehicle. If the classifier indicates that the driver is stressed, fatigued, or anxious, the system  
autonomously switches the vehicle to self-driving mode  
Autonomous Driving Control Module  
The Autonomous Driving Control Module is responsible for taking over the driving functions when the system  
detects that the driver is under stress or in need of assistance.  
Driver Feedback Module  
The Driver Feedback Module ensures that the driver is aware of the system's actions, maintaining a transparent  
interaction between the driver and the vehicle  
Page 2026  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
Data Logging and System Monitoring Module  
The Data Logging and System Monitoring Module monitors the health and performance of the entire system.  
It logs all data from the sensors, classifier outputs, and system statuses, storing them for later analysis  
User Interaction & Control Interface  
The User Interaction & Control Interface provides the driver with a means to manually interact with the  
system. This interface can include a touchscreen display or a voice control system that allows the driver to control  
certain system parameters, such as overriding the automatic mode selection or adjusting comfort preferences.  
Safety and Redundancy Module  
The Safety and Redundancy Module ensures the robustness of the system by implementing backup  
mechanisms and fail-safes. In the event of a sensor malfunction or control unit failure, this module ensures that  
backup sensors or microcontrollers take over to prevent system downtime  
Working  
The system begins when the driver enters the vehicle, and multiple sensors start monitoring the driver's vital  
signs such as pulse rate, oxygen saturation (SpO2), and body temperature. These sensors are continuously  
tracking the physiological data and sending it to the Arduino control unit. The Arduino processes this data,  
filtering it to remove noise and preparing it for further analysis.  
In the Data Processing & Classification Module, the physiological data is fed into a Multi-Layer Perceptron  
(MLP) classifier. The MLP model analyzes the data to determine the driver’s emotional state based on patterns  
learned during training. The model classifies the emotional state into categories such as calm, stressed, or highly  
stressed, If the driver is detected to be stressed or exhibiting signs of fatigue, the system automatically switches  
to autonomous driving mode, taking over the vehicle's control to reduce the risk of human error due to stress.  
In autonomous mode, the vehicle’s Autonomous Driving Control Module activates and takes over all driving  
functions. Using an array of sensors like lidar, cameras, ultrasonic sensors, and GPS, the system navigates the  
vehicle, controlling the steering, acceleration, and braking to safely maneuver through traffic and road  
conditions. It also adjusts the speed based on the environment, ensuring the vehicle moves efficiently and safely  
while the driver can relax and regain composure.  
Throughout the process, the Driver Feedback Module keeps the driver informed of the system’s actions. If the  
system switches to autonomous mode, the driver receives a visual or auditory alert, indicating the transition.  
Additionally, the system might provide reassuring feedback to the driver, especially if stress is detected, helping  
to keep the driver calm.  
Hardware Description  
Pulse Oximeter  
To find the blood oxygen concentration (%), it is first important to know that inside our blood hemoglobin  
is responsible for carrying oxygen. When a person holds a pulse oximeter, light from the device passes  
through the blood in the fingers. This is used to detect the amount of oxygen by measuring the changes in  
light absorption in both oxygenated and deoxygenated blood.  
The MAX30100 sensor consists of two LEDs (Red and IR) and a photodiode. Both of these LEDs are used  
for SPO2 measurement. These two LEDs emit lights at different wavelengths, ~640nm for the red led and  
~940nm for the IR LED. At these particular wavelengths, the oxygenated and deoxygenated hemoglobin  
have vastly different absorption properties.  
Page 2027  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
ATMEGA 328  
ATMEGA 328 microcontroller, which acts as a processor for the arduino board. Nearly it consists of 28 pins.  
From these 28 pins, the inputs can be controlled by transmitting and receiving the inputs to the external device  
Figure: ATMEGA 328  
Analog Input:  
Arduino atmega-328 microcontroller board consist of 6 analog inputs pins. These analog inputs can be named  
from A0 to A5. From these 6 analog inputs pins, we can do the process by using analog inputs. Analog inputs  
can be used in the operating range of 0 to 5V.  
Digital Input:  
Digital inputs are discrete signals represented as 0’s and 1’s. They exist in either an ON or OFF state. The  
Arduino Atmega328 microcontroller has 12 digital pins, D0 to D11. These pins can be used for both input and  
output applications. They trigger and receive discrete pulses, handling only digital inputs.  
IR Sensor  
IR sensor is an electronic device, that emits the light in order to sense some object of the surroundings. An IR  
sensor can measure the heat of an object as well as detects the motion. Usually, in the infrared spectrum, all the  
objects radiate some form of thermal radiation. These types of radiations are invisible to our eyes, but infrared  
sensor can detect these radiations.  
Temperature sensor  
The DHT11 sensor provides digital output for temperature and humidity. It integrates an 8-bit microcontroller  
for reliable performance. Its design ensures high stability and long-term dependability. It uses a resistive element  
and NTC temperature sensing device. The sensor offers fast response, anti-interference ability, and high quality.  
whose mechanism adjusts the speed of the motor, leading them to operate at a certain speed. geared motor have  
the ability to deliver high torque at low speeds, as the gearhead functions as a torque multiplier and can allow  
small motors to generate higher speeds.  
Lithium-ion (Li-ion) battery  
Page 2028  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
A Li-ion battery uses lithium ions in its electrochemistry. During discharge, lithium atoms in the anode ionize  
and release electrons. The ions travel through the electrolyte to the cathode. At the cathode, they recombine with  
electrons and neutralize. Amicro-permeable separator allows ion movement between electrodes. Lithium’s small  
size enables high voltage and charge density. Electrode materials vary across applications. The common pair is  
lithium cobalt oxide (cathode) and graphite (anode). Other cathodes include lithium manganese oxide and  
lithium iron phosphate. Ether compounds are typically used as electrolytes.  
RESULTS AND DISCUSSION  
Table 1 : Comparison  
Model  
Accuracy (%) Precision (%) Recall (%) F1-Score (%)  
Emotion-based Autonomous Control 92.5  
91.2  
86.5  
87.3  
82.7  
89.5  
94.0  
90.2  
91.1  
85.5  
92.0  
92.6  
88.3  
89.2  
84.0  
90.7  
Random Forest (RF)  
88.7  
89.5  
84.1  
91.0  
Support Vector Machine (SVM)  
Logistic Regression (LR)  
Deep Neural Networks (DNN)  
CONCLUSION  
The emotion-based autonomous driving control system successfully integrates real-time physiological  
monitoring with autonomous vehicle technology, providing a more personalized and safe driving experience. By  
continuously assessing the driver’s emotional state through sensors measuring pulse rate, oxygen saturation, and  
body temperature, the system is able to dynamically adjust the vehicle’s driving mode between manual and  
autonomous based on the detected emotional condition. The system’s ability to autonomously take control when  
stress or fatigue is detected helps reduce the risk of human error, particularly in high-stress driving situations.  
Additionally, the real-time feedback provided to the driver ensures transparency, maintaining trust in the system's  
decisions.  
Although the system demonstrated high accuracy in emotional state classification and autonomous driving  
capabilities, there were some limitations regarding external factors influencing sensor readings and minor delays  
in mode transitions at high speeds. These challenges suggest that further optimization of the classification model  
and faster mode-switching algorithms could enhance the system’s reliability, particularly in critical scenarios.  
Despite these limitations, the system has great potential to revolutionize the way we think about autonomous  
driving by considering the emotional well-being of the driver. In future iterations, improvements in sensor  
accuracy, faster response times, and a more robust classifier could make this system even more effective.  
Ultimately, this emotion-based control mechanism not only improves driving safety but also creates a more  
comfortable and adaptive driving environment, marking a significant step forward in human-centered  
autonomous vehicle technology.  
REFERENCES  
1. S. Springer, M. D. Ernst, and A. K. Jain, “Affective human–vehicle interaction: Emotion recognition for  
autonomous driving systems,” IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 3,  
pp. 1452–1465, Mar. 2025.  
2. Q. Qian, Y. Zhang, and L. Wang, “Multi-sensor fusion for autonomous driving: A structured overview,”  
Sensors, vol. 25, no. 2, pp. 112–130, Feb. 2025.  
3. W. Wei, J. Chen, and R. Li, “Fusion techniques for intelligent vehicles: Data, feature, and decision-level  
approaches,” arXiv preprint arXiv:2501.04567, Jan. 2025.  
Page 2029  
www.rsisinternational.org  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026  
4. A. Gupta, R. Sharma, and P. Singh, “Emotion recognition using EEG and facial features for driver state  
monitoring,” IEEE Access, vol. 13, pp. 45210–45222, Apr. 2025.  
5. H. Kim, S. Park, and J. Lee, “IoT-enabled emotion-aware EV systems: Real-time monitoring and adaptive  
driving control,” IEEE Internet of Things Journal, vol. 12, no. 4, pp. 6789–6798, Apr. 2025.  
6. M. Sivaramkrishnan, C. P. Nimmagadda, and V. Manoj, “Smart EV monitoring with IoT and mobile  
integration,” International Journal of Electric and Hybrid Vehicles, vol. 17, no. 1, pp. 33–47, Jan. 2024.  
7. G. M. Fischer, “IoT-based power monitoring and load management for EV charging systems,” IEEE  
Transactions on Smart Grid, vol. 15, no. 2, pp. 987–996, Feb. 2024.  
Page 2030  
www.rsisinternational.org