Exploring the challenges and opportunities of image processing and sensor fusion in autonomous vehicles: A comprehensive review

Abstract

<abstract> <p>Autonomous vehicles are at the forefront of future transportation solutions, but their success hinges on reliable perception. This review paper surveys image processing and sensor fusion techniques vital for ensuring vehicle safety and efficiency. The paper focuses on object detection, recognition, tracking, and scene comprehension via computer vision and machine learning methodologies. In addition, the paper explores challenges within the field, such as robustness in adverse weather conditions, the demand for real-time processing, and the integration of complex sensor data. Furthermore, we examine localization techniques specific to autonomous vehicles. The results show that while substantial progress has been made in each subfield, there are persistent limitations. These include a shortage of comprehensive large-scale testing, the absence of diverse and robust datasets, and occasional inaccuracies in certain studies. These issues impede the seamless deployment of this technology in real-world scenarios. This comprehensive literature review contributes to a deeper understanding of the current state and future directions of image processing and sensor fusion in autonomous vehicles, aiding researchers and practitioners in advancing the development of reliable autonomous driving systems.</p> </abstract>

A Review of Generative Models for 3D Vehicle Wheel Generation and Synthesis

“…The study demonstrated that implementing generative design techniques can enhance wheel performance while minimizing the need for manual intervention in the design process. According to [18], the success of autonomous vehicles, positioned as the future of transportation solutions, relies heavily on reliable perception. Image processing and sensor fusion approaches are thoroughly reviewed in this review paper to guarantee vehicle efficiency and safety.…”

“…To distinguish between different models used for different tasks, Convolutional Neural Networks (CNNs) are vital. With hierarchical data features, CNNs' basic architecture emulates human visual processing abilities, allowing them to recognize, classify, and interpret environmental information [18]. Deep neural networks (DNNs) are commonly utilized for dimensionality reduction in unsupervised learning [33].…”

“…Generative design holds the potential to address the problem of scarce data availability. Within the designer's given limits, designs are independently explored using this computational design process [14], [18], [28], [80], [81]. Recent generative design research combines deep learning and topology optimization to navigate large design areas efficiently [13], [82], [83].…”

“…It navigates across picture segments using tiny filters called kernels and computes dot products based on chromatic information and weights. To create a new cartographic representation, this procedure is applied consistently throughout the image (123 Generative Design and Topology Optimization Place in Product Development Process, n.d.) [10], [11], [18], [55]. The network configuration determines the number of convolutional layers and filter types [49], [84].…”

“…The quality and availability of data sources may impact the depth and reliability or work, limited access to certain datasets or discrepancies in data quality could affect the generalizability of findings for CAD/CAE systems, the rapidly evolving nature of generative modeling techniques and 3D design technologies means that new methodologies or advancements may have emerged since the completion of this review. [18] While the integration of generative models in 3D vehicle wheel design holds promise, practical implementation may encounter various challenges, such as technical constraints, compatibility issues with existing design software, or limitations in computational resources required for complex modeling tasks [19], [20]. Various reviews and research have been done, and despite efforts to maintain objectivity, reviewer biases or subjective interpretations may influence the analysis and synthesis of research findings.…”