International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 01 | Jan 2024
p-ISSN: 2395-0072
www.irjet.net
A REVIEW ON MACHINE LEARNING IN ADAS Vijayalaxmi H1, Aakash A Kumar2,, Arpith M3 ,,Prajwal U4 ,,Vismay UR5 1Assisstant professor, Dept. of Computer Science Engineering, REVA University, Karnataka, India.
2345 UG Student, Dept. of Computer Science Engineering, REVA University, Karnataka, India. ---------------------------------------------------------------------***--------------------------------------------------------------------2.LITREATURE REVIEW Abstract - A Review on Machine Learning in Advanced
Driver Assistance Systems (ADAS)" provides a comprehensive overview of the current state of research in the intersection of machine learning and automotive technology. This review delves into the crucial role that machine learning techniques play in enhancing the capabilities of Advanced Driver Assistance Systems, which are pivotal for improving vehicle safety and automation. The abstract summarizes key advancements in machine learning algorithms and models applied to ADAS, addressing challenges such as object detection, recognition, and decision-making processes in real-time driving scenarios. Additionally, the abstract highlights the evolving landscape of data sources, including sensor fusion and the integration of deep learning methodologies, contributing to the overall efficacy of ADAS. By synthesizing existing literature, this review aims to provide valuable insights into the recent developments, achievements, and potential future directions within the realm of machine learning in ADAS, fostering a deeper understanding of this critical technological domain.
[1] An Intelligent Driving Assistance System Based on Lightweight Deep Learning Model : Published in year 2022 The rise of Advanced Driver Assistance Systems (ADAS) is discussed, emphasizing the use of cameras and radars for environment sensing. The study focuses on the proposed system's use of optical cameras for object detection, utilizing colour features and machine learning techniques. The paper discusses various object detection methods, machine learning technology, and the application of deep learning in the development of the intelligent driving assistance system. The contributions of the study include the collection of comprehensive datasets, the development of a lightweight deep learning model for object recognition, distance estimation method for collision warnings, and situation recognition method for reminding drivers of necessary actions. Brief Insight: The paper discusses the increasing demand for accuracy and efficiency in deep learning models due to technological advancements. It highlights two main approaches for training lightweight deep learning models: compressing pre-training network models and training lightweight network models directly. It specifically mentions Mobile Net and MobileNetV2 as examples of models that use depth-wise convolution in training lightweight network models. The study then applies the lightweight deep learning model for vehicle detection. Overall, the section emphasizes the need for lightweight deep learning models and provides insights into different approaches and technologies used in this area.
Key Words: Object Detection, Lane Detection and Recognition
1.INTRODUCTION The integration of machine learning (ML) techniques in Advanced Driver Assistance Systems (ADAS) has emerged as a transformative force in the automotive industry, revolutionizing the landscape of vehicle safety and automation. As vehicles become increasingly sophisticated, the reliance on intelligent systems capable of interpreting complex driving scenarios and making swift decisions is paramount. This introduction sets the stage for a comprehensive review that explores the synergies between machine learning and ADAS. The discourse navigates through the intricate web of algorithms, models, and applications within this domain, shedding light on the advancements and challenges that define the intersection of machine learning and automotive technology. By delving into the intricate details of ML applications in ADAS, this review aims to provide a holistic understanding of the current state of the field, offering valuable insights into the ongoing evolution and prospects of this dynamic and rapidly evolving technological landscape.
© 2024, IRJET
|
Impact Factor value: 8.226
Concluding, The research paper proposes a computational efficient solution for driver assistance systems that can be implemented on consumer embedded platforms for Advanced Driver Assistance Systems (ADAS). The proposed system includes functionalities such as collision alarm and driver reminding service. Collision alarm is implemented using a lightweight Convolutional Neural Network (CNN) model and distance estimation method, while the driver reminding service is implemented through situation recognition method. Experimental results show that the proposed methods and the adopted model achieve sufficient computational efficiency and performance, particularly with the smaller size of the CNN model, making it more likely to be processed with limited computing resources in embedded systems. In conclusion,
|
ISO 9001:2008 Certified Journal
|
Page 685