The rapid rise of commercial drones has revolutionized urban industries, from package delivery to infrastructure inspection, but their integration into densely populated cities presents significant privacy challenges. In urban environments, where skyscrapers and residential neighborhoods coexist, drones must navigate complex landscapes while avoiding inadvertent surveillance of private spaces. Reports of drones capturing footage of homes have heightened public distrust, with a 2023 study indicating 68% of city residents fear privacy violations. This article explores the multifaceted challenges of urban drone route planning, including the proximity of residential zones, dynamic privacy risks, and regulatory inconsistencies. We review innovative technologies addressing these issues: geofencing establishes no-fly zones over residences but is limited by outdated maps and GPS inaccuracies; AI-powered path optimization prioritizes non-residential routes, though constrained by data quality and computational demands; camera controls, such as automated shutters and real-time image blurring, aim to prevent surveillance but face algorithmic errors and hardware costs; and real-time environmental sensing, using LiDAR and radar, avoids obstacles but struggles with urban interference and high costs. Regulatory frameworks, like FAA and EASA guidelines, and initiatives by various authorities enforce privacy-focused routing. Ethical practices and community engagement will foster public trust and equitable access. Despite significant advancements, limitations like data inaccuracies, cost barriers, and public perception issues persist. As urban air mobility expands, integrating technology, regulation, and ethics will be essential to balance efficiency and privacy, ensuring drones enhance urban life without compromising resident rights.

An in-vehicle network (IVN) is the internal communication network that connects all sensors and control units in an autonomous vehicle. Sensors and control units use the IVN to send perception-related messages and control commands for the normal and safe operation of the vehicle. However, the IVN, by design, is vulnerable to network attacks due to a lack of adequate security mechanisms. This paper presents a Dynamic Windowing Intrusion Detection System (DWIDS) that adapts its detection window in real-time based on observed anomalies, enabling accurate and responsive attack detection. Unlike prior methods that focus on static configurations or single-attack detection, DWIDS supports multi-label classification and real-time tuning of detection parameters. The system is evaluated using two public benchmark datasets (CHD and IVN-IDS challenge) which feature diverse and imbalanced attack types. Experimental results demonstrate high performance across key metrics (e.g., >98% precision, >97% recall and F1-score), including for rare attacks. The findings confirm DWIDS’s practicality and robustness for deployment in real-world autonomous vehicle environments.

This dataset describes RoadSense, a crowdsourcing-enabled data collection framework and accompanying prototype dataset designed to detect and map road surface anomalies using smartphone inertial sensors. The objective is to enable scalable, low-cost road surface condition monitoring by leveraging widely available mobile devices. RoadSense comprises synchronized data from three streams: (i) mobile phone sensors including gyroscope, accelerometer, and GPS; (ii) vehicle telemetry data from the CAN bus via the OBD-II port; and (iii) front-facing video recordings of the road surface. The mobile sensor data is intended for anomaly detection and mapping; video recordings serve as visual ground truth for labeling surface defects such as speed bumps, potholes, and cracks, while CAN bus data provides a supplementary reference for verifying anomalies influenced by vehicle dynamics (e.g., acceleration). The dataset enables reproducible evaluation of road surface monitoring models, facilitates the development of lightweight mobile sensor-based anomaly detection techniques, and supports future fusion with visual models for ground truth validation.

Visual crowd counting has gained serious attention during the last couple of years. The consistent contributions to this topic have now solved several inherited challenges such as scale variations, occlusions, and cross-scene applications. However, these works attempt to improve accuracy and often ignore model size and computational complexity. Several practical applications employ resource-limited stand-alone devices like drones to run crowd models and require real-time inference. Though there have been some good efforts to develop lightweight shallow crowd models offering fast inference time, the relevant literature dedicated to lightweight crowd counting is limited. One possible reason is that lightweight deep-learning models suffer from accuracy degradation in complex scenes due to limited generalization capabilities. This paper addresses this important problem by proposing knowledge distillation to improve the learning capability of lightweight crowd models. Knowledge distillation enables lightweight models to emulate deeper models by distilling the knowledge learned by the deeper model during the training process. The paper presents a detailed experimental analysis with three lightweight crowd models over six benchmark datasets. The results report a clear significance of the proposed method supported by several ablation studies.

Recent advancements have sparked a growing interest in integrating spatiotemporal analysis with large-scale language models. However, spatiotemporal data often contains sensitive information, making it unsuitable for open third-party access. To address this challenge, we propose a Graph-GAN-based model for generating privacy-protected spatiotemporal data. Our approach incorporates spatial and temporal attention blocks in the discriminator and a spatiotemporal deconvolution structure in the generator. These enhancements enable efficient training under Gaussian noise to achieve differential privacy. Extensive experiments conducted on three real-world spatiotemporal datasets validate the efficacy of our model. Our method provides a privacy guarantee while maintaining the data utility. The prediction model trained on our generated data maintains a competitive performance compared to the model trained on the original data.

Severe collisions can result from aggressive driving and poor road conditions, emphasizing the need for effective monitoring to ensure safety. Smartphones, with their array of built-in sensors, offer a practical and affordable solution for road-sensing. However, the lack of reliable, standardized datasets has hindered progress in assessing road conditions and driving patterns. This study addresses this gap by introducing a comprehensive dataset derived from smartphone sensors, which surpasses existing datasets by incorporating a diverse range of sensors including accelerometer, gyroscope, magnetometer, GPS, gravity, orientation, and uncalibrated sensors. These sensors capture extensive parameters such as acceleration force, gravitation, rotation rate, magnetic field strength, and vehicle speed, providing a detailed understanding of road conditions and driving behaviors. The dataset is designed to enhance road safety, infrastructure maintenance, traffic management, and urban planning. By making this dataset available to the community, the study aims to foster collaboration, inspire further research, and facilitate the development of innovative solutions in intelligent transportation systems.

This paper presents Haris, an advanced autonomous mobile robot system for tracking the location of vehicles in crowded car parks using license plate recognition. The system employs simultaneous localization and mapping (SLAM) for autonomous navigation and precise mapping of the parking area, eliminating the need for GPS dependency. In addition, the system utilizes a sophisticated framework using computer vision techniques for object detection and automatic license plate recognition (ALPR) for reading and associating license plate numbers with location data. This information is subsequently synchronized with a back-end service and made accessible to users via a user-friendly mobile app, offering effortless vehicle location and alleviating congestion within the parking facility. The proposed system has the potential to improve the management of short-term large outdoor parking areas in crowded places such as sports stadiums. The demo of the robot can be found on https://youtu.be/ZkTCM35fxa0?si = QjggJuN7M1o3oifx.

Visible Light Communication (VLC) is an emerging technology that has gained significant attention in recent years due to its immense potential for transforming transportation systems. VLC enables short-range high-speed, secure, and reliable vehicle-to-vehicle and vehicle-to-infrastructure communication in various use cases. This work delves into compelling applications of VLC in transportation, highlighting its unique advantages over traditional communication technologies. Leveraging high-speed, secure, and reliable vehicle and vehicle-to-infrastructure communication, VLC offers unprecedented opportunities for enhancing road safety, traffic management, and overall transportation efficiency.

Connected and Autonomous Vehicles (CAVs) are referred to as self-driving vehicles that will become an essential component of future intelligent transportation systems. These CAVs will be equipped with various sensors for perceiving their surroundings and onboard computing capabilities to process sensor data in real-time. Light Detection and Ranging (LiDAR) is one of the essential sensors used for detecting objects and accurate distance estimation. However, LiDAR sensors are susceptible to several types of attacks. Adversaries can exploit LiDAR sensors either physically, by sending signals directly to the sensor, or digitally, by manipulating LiDAR data after gaining access to the in-vehicle network. Over the past few years, there has been significant research on the vulnerabilities, attack models, and security of LiDAR sensors. However, to our knowledge, no comprehensive survey exists that addresses these aspects of autonomous vehicle security. This paper aims to bridge this gap by presenting an overview of LiDAR-based perception, data processing, threat models, and defense mechanisms for LiDAR sensors in CAVs. We believe this paper will serve as a valuable reference for researchers, providing a clear understanding of cyber-physical attacks and defense strategies related to LiDAR sensors in autonomous vehicles and related fields.

Most state-of-the-art crowd counting methods use color (RGB) images to learn the density map of the crowd. However, these methods often struggle to achieve higher accuracy in densely crowded scenes with poor illumination. Recently, some studies have reported improvement in the accuracy of crowd counting models using a combination of RGB and thermal images. Although multimodal data can lead to better predictions, multimodal data might not be always available beforehand. In this paper, we propose the use of generative adversarial networks (GANs) to automatically generate thermal infrared (TIR) images from color (RGB) images and use both to train crowd counting models to achieve higher accuracy. We use a Pix2Pix GAN network first to translate RGB images to TIR images. Our experiments on several state-of-the-art crowd counting models and benchmark crowd datasets report significant improvement in accuracy.

The success of modern deep learning is attributed to two key elements: huge amounts of training data and large model sizes. Where a vast amount of data allow the model to learn more features, the large model architecture boosts the learning capability of the model. However, both these factors result in prolonged training time. In some practical applications such as edge-based learning and federated learning, limited-time budgets necessitate more efficient training methods. This paper proposes an effective technique for training arbitrary deep learning models within fixed time constraints utilizing sample importance and dynamic ranking. The proposed method is extensively evaluated in both classification and regression tasks in computer vision. The results consistently show clear gains achieved by the proposed method in improving the learning performance of various state-of-the-art deep learning models in both regression and classification tasks.

Recent advances in deep learning techniques have achieved remarkable performance in several computer vision problems. A notably intuitive technique called Curriculum Learning (CL) has been introduced recently for training deep learning models. Surprisingly, curriculum learning achieves significantly improved results in some tasks but marginal or no improvement in others. Hence, there is still a debate about its adoption as a standard method to train supervised learning models. In this work, we investigate the impact of curriculum learning in crowd counting using the density estimation method. We performed detailed investigations by conducting 112 experiments using six different CL settings using eight different crowd models. Our experiments show that curriculum learning improves the model learning performance and shortens the convergence time.

The rapid evolution of modern automobiles into intelligent and interconnected entities presents new challenges in cybersecurity, particularly in Intrusion Detection Systems (IDS) for In-Vehicle Networks (IVNs). This survey paper offers an in-depth examination of advanced machine learning (ML) and deep learning (DL) approaches employed in developing sophisticated IDS for safeguarding IVNs against potential cyber-attacks. Specifically, we focus on the Controller Area Network (CAN) protocol, which is prevalent in in-vehicle communication systems, yet exhibits inherent security vulnerabilities. We propose a novel taxonomy categorizing IDS techniques into conventional ML, DL, and hybrid models, highlighting their applicability in detecting and mitigating various cyber threats, including spoofing, eavesdropping, and denial-of-service attacks. We highlight the transition from traditional signature-based to anomaly-based detection methods, emphasizing the significant advantages of AI-driven approaches in identifying novel and sophisticated intrusions. Our systematic review covers a range of AI algorithms, including traditional ML, and advanced neural network models, such as Transformers, illustrating their effectiveness in IDS applications within IVNs. Additionally, we explore emerging technologies, such as Federated Learning (FL) and Transfer Learning, to enhance the robustness and adaptability of IDS solutions. Based on our thorough analysis, we identify key limitations in current methodologies and propose potential paths for future research, focusing on integrating real-time data analysis, cross-layer security measures, and collaborative IDS frameworks.

Autonomous vehicles (AVs) are the cornerstone of the future intelligent transportation systems. The AVs are intelligent vehicles with sophisticated sensing capabilities powered by advanced artificial intelligence. Among the various sensors, cameras constitute the most significant sensor that enables the AV to perceive the environment in real time to safely navigate. However, the safety of AVs depends upon the accuracy of object detection using cameras, and any attack on the camera may cause incidents. This paper demonstrates how a simple hardware setup can enable spoofing attacks on the AV camera to inject fake objects into the camera video feed to fool the AV perception system. Our experiments show that such attacks can be deployed physically by attaching a simple hardware setup and the attacker can spoof objects at any instance without accessing the camera feed. Lastly, the paper provides insights on how to mitigate such attacks in AVs.

Deep learning models require an enormous amount of data for training. However, recently there is a shift in machine learning from model-centric to data-centric approaches. In data-centric approaches, the focus is to refine and improve the quality of the data to improve the learning performance of the models rather than redesigning model architectures. In this paper, we propose CLIP i.e., Curriculum Learning with Iterative data Pruning. CLIP combines two data-centric approaches i.e., curriculum learning and dataset pruning to improve the model learning accuracy and convergence speed. The proposed scheme applies loss-aware dataset pruning to iteratively remove the least significant samples and progressively reduces the size of the effective dataset in the curriculum learning training. Extensive experiments performed on crowd density estimation models validate the notion behind combining the two approaches by reducing the convergence time and improving generalization. To our knowledge, the idea of data pruning as an embedded process in curriculum learning is novel.

Crowd Management (CM) is an important public safety paradigm that monitors, manages, and even guides a group of people to ensure their safety and prevent any unpredicted disturbance. Recent CM approaches utilize IoT devices such as CCTVs and RFID tags to enable crowd monitoring and tracking as well as abnormal behavior detection and prediction. However, these systems still suffer from several short-comings, such as limited mobility and poor response time. In this magazine, we overview the potential of CM platforms supported by an aerial infrastructure composed of a fleet of Unmanned Aerial Vehicles (UAVs). Thanks to their mobility, sky vision, and flexibility, UAVs can provide adaptive CM services and intervene immediately to assess the crowd status. The article discusses the challenges of designing inno-vative solutions to enhance the autonomy of UAVs and their use in real-world applications. Finally, future research directions are pointed out to develop futuristic UAV-based CM systems.

Visual crowd counting estimates the density of the crowd using deep learning models such as convolution neural networks (CNNs). The performance of the model heavily relies on the quality of the training data that constitutes crowd images. In harsh weather such as fog, dust, and low light conditions, the inference performance may severely degrade on the noisy and blur images. In this paper, we propose the use of Pix2Pix generative adversarial network (GAN) to first denoise the crowd images prior to passing them to the counting model. A Pix2Pix network is trained using synthetic noisy images generated from original crowd images and then the pretrained generator is then used in the inference engine to estimate the crowd density in unseen, noisy crowd images. The performance is tested on JHU-Crowd dataset to validate the significance of the proposed method particularly when high reliability and accuracy are required.

Automatic crowd counting using density estimation has gained significant attention in computer vision research. As a result, a large number of crowd counting and density estimation models using convolution neural networks (CNN) have been published in the last few years. These models have achieved good accuracy over benchmark datasets. However, attempts to improve the accuracy often lead to higher complexity in these models. In real-time video surveillance applications using drones with limited computing resources, deep models incur intolerable higher inference delay. In this paper, we propose (i) a Lightweight Crowd Density estimation model (LCDnet) for real-time video surveillance, and (ii) an improved training method using curriculum learning (CL). LCDnet is trained using CL and evaluated over two benchmark datasets i.e., DroneRGBT and CARPK. Results are compared with existing crowd models. Our evaluation shows that the LCDnet achieves a reasonably good accuracy while significantly reducing the inference time and memory requirement and thus can be deployed over edge devices with very limited computing resources.

Crowd counting is an effective tool for situational awareness in public places. Automated crowd counting using images and videos is an interesting yet challenging problem that has gained significant attention in computer vision. Over the past few years, various deep learning methods have been developed to achieve state-of-the-art performance. The methods evolved over time vary in many aspects such as model architecture, input pipeline, learning paradigm, computational complexity, and accuracy gains etc. In this paper, we present a systematic and comprehensive review of the most significant contributions in the area of crowd counting. Although few surveys exist on the topic, our survey is most up-to date and different in several aspects. First, it provides a more meaningful categorization of the most significant contributions by model architectures, learning methods (i.e., loss functions), and evaluation methods (i.e., evaluation metrics). We chose prominent and distinct works and excluded similar works. We also sort the well-known crowd counting models by their performance over benchmark datasets. We believe that this survey can be a good resource for novice researchers to understand the progressive developments and contributions over time and the current state-of-the-art.