Search

The challenging problem in the traffic system is lane detection. This Lane detection which attracts the computer vision community’s attention. For computer vision and machine learning techniques, the Lane detection which acts as the multi-feature detection problem. Many machine learning techniques are used for lane detection. Driver support system is one of the most important features in the modern vehicles to ensure the safety of the driver and decrease the vehicle accidents on road. Road Lane detection is the most challenging task and complex tasks now-a-days. Road localization and relative position between vehicle and roads which also includes with this. But in this journal, we propose a new method. Here, an on- board camera to be used which is looking outwards are presented here in this work. This proposed technique which can be used for all types of roads like painted, unpainted, curved, straight roads etc in different weather conditions. No need for camera calibration and coordination of the transform, may be any changing illumination situation, shadow effects, various road types. No representation for speed limits. This includes that the system acquires the front view using a camera mounted on the vehicles and detect the Lane by applying the code from the Python Programming process. This proposed system does not require any more information about roads. This system which demonstrates a robust performance for Lane detection.

Driver Assistance Systems (DAS) have been progressively incorporated into commercial vehicles in recent years. All these systems are paving the way for the forthcoming autonomous vehicle which will become a reality in the near future. Existing systems are based on numerous electronic systems with advanced skills, high performances, and high degrees of adaptability and intelligence. As is to be expected, these cutting-edge features require, in most cases, the use of powerful computing platforms. However, the deployment of such platforms is not an easy task, since they have to be integrated in the vehicle where there exist important restrictions regarding size, power consumption and cost. In this sense, every smart proposal aimed at reducing the complexity of these systems without degrading performance, is always a valuable contribution in the field. In this work, we propose a methodology to reduce the dimensionality of a driver distraction recognition system. The methodology is based on a multi-objective genetic algorithm that looks for the minimum set of useful features collected during the driving task and also for the simplest recognition system. The recognition algorithm is an Extreme Learning Machine (ELM) whose simplicity and fast learning procedure make it especially suitable to be used by a Genetic Algorithm which needs to evaluate thousands of candidate solutions. The proposed methodology has been tested with a real-world database collected from different drivers performing an itinerary with an instrumented car. The results obtained validate the proposal as a method to reduce the complexity of a driver distraction recognition system.

Intelligent systems, like driver assistance systems, remain within vehicles and help drivers by providing essential information about traffic, blockage of roads, and possible routes for safe driving. The objective of scene text spotting in a driver assistance system is to localize and recognize scene texts, signs of milestones, traffic panels, and road marks in natural scene images. However, text edges get fainted due to adverse weather conditions, like fog, rain, smog, or poor contrast. This makes the task of spotting more challenging. In this paper, we propose an end-to-end trainable deep neural network, known as Met-MLTS, that can address the issue of spotting multi-oriented text instances in scene images captured in adverse meteorological conditions. It localizes words, predicts script class, and performs word spotting for every rotated bounding box. It is a fast multilingual scene text spotter that utilizes hierarchical spatial context, channel-wise inter-dependencies, and semantic edge supervision to localize and recognize words and predict script class in scene images using smartphones. We explore inter-class interference to reduce the misclassification problem. A light-weight recognition module for multilingual character segmentation, word-level recognition, and script identification is incorporated. We demonstrate the efficacy of our spotting network on resource-constraint devices.

As automobiles have become an essential part to facilitate our daily life, Advanced Driver Assistance Systems (ADAS) have been gaining more and more interest in assisting drivers to enhance both safety and convenience. To respond timely in case of an emergency, ADAS needs to keep track of the driver’s health/consciousness, which is generally achieved by monitoring the driver’s vital signs including Respiration Rate (RR), Heart Rate (HR) and Heart Rate Variability (HRV). However, most of the state-of-art solutions need to assume that the human is stationary, which does not hold in practical driving scenarios. To tackle the problem, we propose a novel system, which can estimate driver’s RR, HR and Inter-Beat Intervals (IBI) in the presence of driver’s motion artifacts using commercial millimeter-wave (mmWave) radio. The system consists of two key components. First, to extract the reflection signals containing vital signals, the motion artifacts are first removed by a novel motion compensation module, followed by the periodicity check to identify the components with vital signals. Second, the respiration and heartbeat signals are reconstructed by jointly optimizing the decomposition of all the extracted compound vital signals over different range-azimuth bins. We evaluate the system performance in real driving environment and investigate the impact of different parameters including the device locations, pavement conditions and motion types. Experimental results show that the proposed system can achieve a median error of 0.16 Respiration Per Minute (RPM), 0.82 Beat Per Minute (BPM) and 46 ms for RR, HR and IBI estimations, corresponding to the relative accuracy of 99.17%, 98.94% and 94.11%, respectively.

Lane detection is one of the indispensable and key elements of self-driving environmental perception. Many lane detection models have been proposed, solving lane detection under challenging conditions, including intersection merging and splitting, curves, boundaries, occlusions and combinations of scene types. Nevertheless, lane detection will remain an open problem for some time to come. The ability to cope well with those challenging scenes impacts greatly the applications of lane detection on advanced driver assistance systems (ADASs). In this paper, a spatio-temporal network with double Convolutional Gated Recurrent Units (ConvGRUs) is proposed to address lane detection in challenging scenes. Both of ConvGRUs have the same structures, but different locations and functions in our network. One is used to extract the information of the most likely low-level features of lane markings. The extracted features are input into the next layer of the end-to-end network after concatenating them with the outputs of some blocks. The other one takes some continuous frames as its input to process the spatio-temporal driving information. Extensive experiments on the large-scale TuSimple lane marking challenge dataset and Unsupervised LLAMAS dataset demonstrate that the proposed model can effectively detect lanes in the challenging driving scenes. Our model can outperform the state-of-the-art lane detection models., Comment: 13 pages, 9 figures, 8 tables, Accepted by IEEE Transactions on Intelligent Transportation Systems

Advanced driver assistance systems have been an active research topic for decades, for which many approaches have been developed not only to reduce the number of traffic accidents but also to increase the driver's comfort. Among the many different solutions proposed, learning-based prediction approaches have gained considerable attention in recent years. Within this scope, this work focuses on the implementation aspects of a linear learning-based regression model for detecting unintended lane-departures, where the goal is to achieve a prediction model with good predictive performance while keeping the computational complexity as low as possible. Aspects under consideration include input signal selection and down-sampling. The linear prediction model is analyzed using a real world data set, and benchmarked against a kinematic constant velocity model and a non-linear regression model. The results show that the linear regression model has a significantly higher prediction performance when compared to a kinematic model. It is also shown that the predictive performance remains comparable to the more complex nonlinear regression model, even though the computational complexity of the linear model is significantly lower.

Vision-based intelligent systems such as driver assistance systems and transportation systems should take into account weather conditions. The presence of haze in images can be a critical threat to driving scenarios. Haze density measures the visibility and usability of hazy images captured in real-world conditions. The prediction of haze density can be valuable in various vision-based intelligent systems, especially in those systems deployed in outdoor environments. Haze density prediction is a challenging task since the haze and many scene contents have a lot in common in appearance. Existing methods generally utilize different priors and design complex handcrafted features to predict the visibility or haze density of the image. In this article, we propose a novel end-to-end convolutional neural network (CNN) based method to predict haze density, named as HazDesNet. Our HazDesNet takes a hazy image as input and predicts a pixel-level haze density map. The density map is then refined and smoothed, and the average of the refined map is calculated as the global haze density of the image. To verify the performance of HazDesNet, a subjective human study is performed to build a Human Perceptual Haze Density (HPHD) database, which includes 500 real-world hazy images and 100 synthetic hazy images, and the corresponding human-rated perceptual haze density scores. Experimental results show that our method achieves the best haze density prediction performance on our built HPHD database and existing databases. Besides the global quantitative results, our HazDesNet is capable of predicting a continuous, stable, fine, and high-resolution haze density map. We will make the database and code publicly available at https://github.com/JiaheZhang/HazDesNet.

Current research and development in recognizing and predicting driving behaviors plays an important role in the development of Advanced Driver Assistance Systems (ADAS). For this reason, many machine learning approaches have been developed and applied. Hidden Markov Model (HMM) is a suitable algorithm due to its ability to handle time series data and state transition descriptions. Therefore, this contribution will focus on a review of HMM and its applications. The aim of this contribution is to analyze the current state of various driving behavior models and related HMM-based algorithms. By examining the current available approaches, a review is provided with respect to: i) influencing factors of driving behaviors corresponding to the research objectives of different driving models, ii) summarizing HMM related methods applied to driving behavior studies, and iii) discussing limitations, issues, and future potential works of the HMM-based algorithms. Conclusions with respect to the development of intelligent driving assistant system and vehicle dynamics control systems are given.

The last decade has witnessed a growing demand for precise positioning in many applications, including autonomous car navigation. The safety features in autonomous driving and Advanced Driver Assistance Systems (ADAS) require lane-level positioning accuracy. Such accuracy can be obtained from the Global Navigation Satellite Systems (GNSS) through either differential techniques or Precise Point Positioning (PPP). PPP is currently favored over differential GNSS because it provides a global solution without the need for local reference stations. Nevertheless, employing PPP for land vehicles would be challenging due to frequent signal degradation and blockage. Integrating PPP with an Inertial Navigation System (INS) can solve the solution continuity problem; however, the INS solution drifts over time, resulting in losing the desired accuracy. Implementing a reliable PPP/INS system that can preserve the required accuracy is not trivial, especially with financial and computational cost constraints. This article proposes the integration of PPP with the Reduced Inertial Sensor System (RISS) for lane-level car navigation. The high-precision needed in lane-level positioning can be achieved by integrating PPP with high-end INS. Since high-end INS are expensive, this work proposes the use of RISS instead of the traditional INS. RISS uses only one gyroscope and two accelerometers, which can save more than half the high-end INS cost. The proposed PPP/RISS system was tested through three road tests that included highway driving under several overpasses. The system was able to maintain horizontal position errors of less than 50 cm.

In the field of advanced driver assistance systems (ADAS), effective learning of driver fatigue characteristics representation is a major challenge due to uncertainties of both real roads and drivers. To tackle this problem, this paper proposes a novel model of learning interpretable representations for fatigue features, so as to improve the performance of the monitoring system on real roads through the learned features. First, the approximate entropy of the Steering Wheel Angle SWAsequence is used to crop adaptive lengths of the Recurrent Neural Network (RNN) input data. Then, it will learn statistical indexes to discover the random steering characteristics. The Long Short-Term Memory (LSTM) unit can memorize drivers long-term operating characteristics and instantaneous change patterns, and mine their fatigue characteristics. Finally, the information gain method is used to discover the strong correlation between potential characteristics and fatigue levels, thereby obtaining the best feature representation for driver fatigue. The proposed method makes use of the advantages of recurrent neural network learning to explicitly capture various potential characteristics of fatigue driving and interactions between non-linear characteristics in driving. Experimental results on real driving data validate the effectiveness of our proposed method.

A major challenges of deep learning (DL) is the necessity to collect huge amounts of training data. Often, the lack of a sufficiently large dataset discourages the use of DL in certain applications. Typically, acquiring the required amounts of data costs considerable time, material and effort. To mitigate this problem, the use of synthetic images combined with real data is a popular approach, widely adopted in the scientific community to effectively train various detectors. In this study, we examined the potential of synthetic data-based training in the field of intelligent transportation systems. Our focus is on camera-based traffic sign recognition applications for advanced driver assistance systems and autonomous driving. The proposed augmentation pipeline of synthetic datasets includes novel augmentation processes such as structured shadows and gaussian specular highlights. A well-known DL model was trained with different datasets to compare the performance of synthetic and real image-based trained models. Additionally, a new, detailed method to objectively compare these models is proposed. Synthetic images are generated using a semi-supervised errors-guide method which is also described. Our experiments showed that a synthetic image-based approach outperforms in most cases real image-based training when applied to cross-domain test datasets (+10% precision for GTSRB dataset) and consequently, the generalization of the model is improved decreasing the cost of acquiring images.

Many intelligent transportation systems, such as advanced driver assistance systems (ADAS), have been being developed to increase traffic safety. ADAS have the potential to save lives and reduce crashes by eliminating the human error in driving process. The Lane detection system is designed to identify and estimate the position of lane boundaries in front of the ego-vehicle. Thus, it serves as a crucial and fundamental component in various ADAS, for instance, lane keeping assistance and lane departure warning assistance systems. In this paper, a proposed method of vision-based lane detection in urban roads is presented. First, we define a region of interest to exclude the misleading parts of the road image, then a bird’s-eye view of the road in front of the ego-vehicle is obtained by employing the inverse perspective mapping. Second, we utilize the distinct colors of lane markings to achieve robust lane markings candidate detection. Finally, the estimated lane boundaries are represented by quadratic models whose parameters are estimated from the detected lane pixels using the RANSAC algorithm. Furthermore, we present a thorough evaluation of the performance of the proposed method using the ground-truth data of the Caltech dataset and a comparative analysis between the quadratic model used and other models presented in the literature. Detection results show the effectiveness of the proposed method in detecting lane boundaries in different conditions in urban roads, including curved lanes, shadows, illumination variations and presence of street writings. Moreover, the overall process takes an average time of 30.63 milliseconds per frame.

Deep Neural Networks (DNNs) are increasingly applied in the real world in safety critical applications like advanced driver assistance systems. An example for such use case is represented by traffic sign recognition systems. At the same time, it is known that current DNNs can be fooled by adversarial attacks, which raises safety concerns if those attacks can be applied under realistic conditions. In this work we apply different black-box attack methods to generate perturbations that are applied in the physical environment and can be used to fool systems under different environmental conditions. To the best of our knowledge we are the first to combine a general framework for physical attacks with different black-box attack methods and study the impact of the different methods on the success rate of the attack under the same setting. We show that reliable physical adversarial attacks can be performed with different methods and that it is also possible to reduce the perceptibility of the resulting perturbations. The findings highlight the need for viable defenses of a DNN even in the black-box case, but at the same time form the basis for securing a DNN with methods like adversarial training which utilizes adversarial attacks to augment the original training data.

Among the current advanced driver assistance systems, Vehicle-to-Vehicle (V2V) technology has great potential to increase Vehicular Ad Hoc Network (VANET) performance in terms of security, energy efficiency, and comfortable driving. In reality, vehicle drivers regularly change lanes depending on their assumptions regarding visual distances. However, many systems are not quite well-designed, because the visible range is limited, making it difficult to achieve such a task. V2V technology offers high potential for VANET to increase safety, energy efficiency, and driver convenience. Drivers can make more intelligent options in terms of lane selection using predicted information of downstream lane traffic, which is essential for obtaining mobility benefits., Comment: 6 pages

As a key environment perception technology of autonomous driving or driver assistance systems, lane detection is to ensure vehicles to drive safely in corresponding lane. However, existing lane detection algorithms for two-lane detection focus on using various filtering methods to reduce the impact of useless information, resulting in low accuracy and low efficiency. In this paper, a novel Adaptive Region of Interest (A-ROI) extraction method is proposed to improve the accuracy and real-time performance of the two-lane detection algorithm. Three key technologies are introduced to solve the problems. First, A-ROI, which only focuses on the lane where the vehicle is located, is applied to the Bird’s-Eye-View image obtained by using Inverse Perspective Mapping (IPM). Next, based on Bayesian framework and Likelihood models, a lane feature extraction method with a lane-like feature filter is used for edge detection. Finally, an improved Random Sample Consensus (RANSAC) algorithm is introduced by using a filter that can remove noisy lane data. The performance of the proposed A-ROI method together with the improved lane detection method is evaluated via simulation of various scenarios. Experimental results show the proposed method has better accuracy and real-time performance than the traditional lane detection methods.

Driver gaze zone estimation is an important task in Advanced Driver Assistance Systems (ADAS), which suffers difficulties including head pose, capture direction, glass occlusion, and real-time requirement, etc. Most previous methods combine face modalities and head pose using concat process, which may result in over-fitting due to the unbalanced dimension. Focusing on gaze zone estimation problems, we propose the Head Pose Fusion Assisted supervision & Eye Region Weighted Encoding (HP-ERW) structure, which fuses head pose attribute and face modalities together through spatial attention and Kronecker product mechanisms. Firstly, we introduce a pre-processing module dealing with head pose and face information, with the purpose of extracting input vectors and improving the fusion speed of the HP-ERW structure. Secondly, an Eye Region Weighted Encoding Network (ERW-Net) based on spatial attention is proposed to strengthen the networks perception ability for encoding features. Finally, we propose a dual-channel Head Pose Fusion Network (HP-Net) based on the Kronecker product mechanism, with the purpose of fusing head pose and improving the estimation accuracy. Experiments show that the HP-ERW outperforms compared existing methods on several public datasets. The designed ADAS using the proposed method achieves 23.5 fps real-time application with small memory requirement of 4,884 KB.

Stereo matching aiming to perceive the 3-D geometry of a scene facilitates numerous computer vision tasks used in advanced driver assistance systems (ADAS). Although numerous methods have been proposed for this task by leveraging deep convolutional neural networks (CNNs), stereo matching still remains an unsolved problem due to its inherent matching ambiguities. To overcome these limitations, we present a method for jointly estimating disparity and confidence from stereo image pairs through deep networks. We accomplish this through a minmax optimization to learn the generative cost aggregation networks and discriminative confidence estimation networks in an adversarial manner. Concretely, the generative cost aggregation networks are trained to accurately generate disparities at both confident and unconfident pixels from an input matching cost that are indistinguishable by the discriminative confidence estimation networks, while the discriminative confidence estimation networks are trained to distinguish the confident and unconfident disparities. In addition, to fully exploit complementary information of matching cost, disparity, and color image in confidence estimation, we present a dynamic fusion module. Experimental results show that this model outperforms the state-of-the-art methods on various benchmarks including real driving scenes.

As advanced driver assistance systems (ADAS) increase in functionality, so do the functional requirements of their visual displays to communicate status and settings. Drivers who misinterpret in-vehicle displays may be more likely to misuse or avoid using ADAS, negating the potential safety benefits. A novel evaluation method is proposed and applied to the (at the time of surveying) latest Volvo instrument cluster. An online survey was completed by 838 Volvo car owners with varying levels of familiarity with adaptive cruise control and pilot assist. Results suggest that the concepts of a system being “available but not on” or “on but not actively providing support” can be difficult for novice users to recognize, resulting in a mode or state confusion. However, more experienced users were less likely to make these types of errors. The assessment method of providing users with partially masked images shows promise as a useful tool for evaluating comprehension of ADAS in-vehicle visual displays.

Road markings are beneficial to human drivers, advanced driver assistance systems (ADAS), and automated driving systems (ADS), on the contrary, snow coverage on roads poses a challenge to all three of these groups with respect to lane detection, as white road markings are difficult to distinguish from snow. Indeed, yellow road markings provide a visual contrast to snow that can increase a human drivers’ visibility. Yet, in spite of this fact, yellow road markings are becoming increasingly rare in Europe due to the high costs of painting and maintaining two road marking colors. More importantly, in conjunction with our increased reliance on automated driving, the question of whether yellow road markings are of value to automatic lane detection functions arises. To answer this question, images from snowy conditions are assessed to see how different representations of colors in images (color spaces) affect the visibility levels of white and yellow road markings. The results presented in this paper suggest that yellow markings provide a certain number of benefits for automated driving, offering recommendations as to what the most appropriate color spaces are for detecting lanes in snowy conditions. To obtain the safest and most cost-efficient roads in the future, both human and automated drivers’ actions must be considered. Road authorities and car manufacturers also have a shared interest in discovering how road infrastructure design, including road marking, can be adapted to support automated driving.

Advanced driver-assistance systems need a camera calibration algorithm for various vision applications including surround-view monitoring (SVM) and lane departure warning (LDW). Although cameras mounted on a vehicle are calibrated in the manufacturing process, their orientation angles are subject to tilting because of continuing vibration and external impact. To solve the problem, this paper presents an online calibration algorithm for camera orientation estimation using motion vectors and three-dimensional geometry. The proposed algorithm consists of three steps: i) driving direction estimation by calculating an intersection of motion vectors, ii) camera orientation estimation based on 3-line random sample consensus (RANSAC) using the estimated intersection, and iii) final orientation decision using extended Kalman filter from the result of each frame. Experimental results demonstrate that the proposed algorithm stably estimates camera orientation angles from motion vectors and lines under the parallelism and orthogonality assumptions.

Machine Learning (ML) involves making a machine able to learn and take decisions on real-life problems by working with an efficient set of algorithms. The generated ML models find application in different areas of research and management. One such field, automotive technology, employs ML enabled commercialized advanced driver assistance systems (ADAS) which include traffic sign recognition as a part. With the increasing demand for the intelligence of vehicles, and the advent of self-driving cars, it is extremely necessary to detect and recognize traffic signs automatically through computer technology. For this, neural networks can be applied for analyzing images of traffic signs for cognitive decision making by autonomous vehicles. Neural networks are the computing systems which act as a means of performing ML. In this work, a convolutional neural network (CNN) based ML model is built for recognition of traffic signs accurately for decision making, when installed in driverless vehicles.

Logs of vehicle controller area network (CAN) bus traffic supplemented by accelerometer and GPS data can provide valuable information about the use and operation of advanced driver assistance systems (ADAS) to the broader safety research community. Although CAN bus message codes are often manufacturer-specific, third-party libraries provide partial decoding of messages from many vehicle models, which can be augmented by reverse-engineering additional signals. This study explored the value of CAN bus, accelerometer, and GPS data that were logged on a variety of light vehicle models with an emphasis on availability of lane keep assist/lane departure warning, and adaptive cruise control. This study demonstrated that in-vehicle ADAS variables such as system on/off status and whether the system is actively controlling the vehicle could be determined on a variety of vehicle types. Associated control variables such as steering wheel angle, gas pedal state, and brake pedal state could also be determined from most vehicles tested. CAN messages, together with roadway features identified via GPS location, can provide a richer understanding of ADAS efficacy. Comparisons of message structure between models may also inform standardization efforts for electronic data recorders and telematics.

Advanced Driver Assistance Systems (ADAS) provide safety and comfort while driving. However, to effectively use ADAS, it is necessary for users to have proper knowledge of the systems and to trust the system to operate safely. Providing knowledge about operational capabilities and limitations of a system may help improve drivers’ mental models and calibrate their trust resulting in proper use of ADAS. Traditionally system information is provided via the owner’s manual, which is known to be tedious and time-consuming and underscores the need for alternate training approaches. This study evaluates two training methods, Text-Based and System Visualization, to examine users’ perceptions of training and change in trust after training. Results show that although training did not affect users’ trust, a qualitative examination showed that users preferred the Text-Based method rather than the Visualization method.

Contemporary transportation is developing toward applications of large-capacity vehicles. This trend is caused mainly by environmental and economic factors. In the public transportation area, it leads to the use of long and articulated electric buses, which are difficult to maneuver, even for experienced drivers. Consequently, there is a need to develop advanced driver-assistance systems (ADAS) for operators of urban buses to safely perform positioning maneuvers with acceptable precision. This article presents an ADAS concept dedicated to the task of precisely docking a pantograph tip with a charging station. We describe the functional structure of the ADAS and explain details of its interconnected components responsible for motion algorithmization, environmental perception, and vehicle localization. Experimental results of assisted pantograph docking maneuvers obtained with an articulated electric bus are provided.

The emergence of vehicle technologies that promote driver safety and convenience calls for investigation of the prevalence of driver assistance systems as well as of their use rates. A consumer driven understanding as to why certain vehicle technology is used remains largely unexplored. We examined drivers’ experience using 13 different advanced driver assistance systems (ADAS) and several reasons that may explain rates of use through a nationally-distributed survey. Our analysis focused on drivers’ levels of understanding and trust with their vehicle’s ADAS as well as drivers’ perceived ease, or difficulty, in using the systems. Respondents’ age and experience with Level 0 or Level 1 technologies revealed additional group differences, suggesting older drivers (55+), and those with only Level 0 systems as using ADAS more often. These data are interpreted using the Driver Behavior Questionnaire framework and offer a snapshot of the pervasiveness of certain driver safety systems.

In order to test and verify radar-based systems for advanced driver assistance systems, radar target simulators offer a test environment for reproducible dynamic automotive scenarios. This article presents a system capable of generating multiple targets with individual ranges and velocities for chirp-sequence frequency modulated continuous waveform radars. The novel approach allows to generate targets without prior knowledge about the radar sensor by estimating the radar’s waveform parameters. The proposed simulator offers high flexibility at low costs by performing the estimation on undersampled data. Finally, an automotive scenario of a spatially extended target is simulated for a radar sensor with the proposed system.

Assessment methods are needed to rate the performances of advanced driver assistance systems in a range of real-world conditions, enabling the possibility of mandating minimum performance requirements beyond standardized, regulatory pass-or-fail tests, and ultimately ensuring a real and objectively measurable safety benefit. To bridge the gap between regulatory and real-world performance, this work presents a novel robustness assessment methodology and defines a robustness index determined from regulatory tests to analyze the real-world performance of lane departure warning (LDW) systems. In this context, a robust system means that it is insensitive to changes in driving conditions or environmental conditions. Distance to line (DTL) and time to line crossing (TTLC) were calculated for a light truck and a passenger car, and a black box model of the LDW systems was developed to predict their performance over different lane markings, drifting directions, and vehicle lateral and longitudinal speeds. During the test, neither of the vehicles triggered warning in around 10% of the trials despite the perfect condition of the markings painted on the proving ground. The type of lane marking significantly influenced DTL for both vehicles. For the light truck, the drifting direction, marking type, and their interaction were found to be statistically significant, which resulted in a lower robustness index than that of the passenger car. For both vehicles, TTLC was inversely proportional to the lateral speed, which greatly influences crash avoidance.

In recent years, autonomous vehicle driving technology and advanced driver assistance systems have played a key role in improving road safety. However, weather conditions such as snow pose severe challenges for autonomous driving and are an active research area. Thanks to their superior reliability, the resilience of detection, and improved accuracy, advances in computation and sensor technology have paved the way for deep learning and neural network–based techniques that can replace the classical approaches. In this research, we investigate the semantic segmentation of roads in snowy environments. We propose a multi-modal fused RGB-T semantic segmentation utilizing a color (RGB) image and thermal map (T) as inputs for the network. This paper introduces a novel fusion module that combines the feature map from both inputs. We evaluate the proposed model on a new snow dataset that we collected and on other publicly available datasets. The segmentation results show that the proposed fused RGB-T input can segregate human subjects in snowy environments better than an RGB-only input. The fusion module plays a vital role in improving the efficiency of multiple input neural networks for person detection. Our results show that the proposed network can generate a higher success rate than other state-of-the-art networks. The combination of our fused module and pyramid supervision path generated the best results in both mean accuracy and mean intersection over union in every dataset.

Advanced driver assistance systems with SAE Level 2 automated capabilities have entered the vehicle marketplace. Such driving automation systems (DASs) have the potential to fundamentally change the driving experience through automated lateral and longitudinal vehicle control. However, drivers may not use DASs as intended because of their misunderstanding of the systems’ capabilities and limitations. Moreover, the real-world use and effects of this novel technology on transportation safety are largely unknown. To investigate driver interactions with driving automation, the study examined existing naturalistic driving data collected from 50 participants who drove personally owned vehicles with Level 2 DASs for 12 months. It was found that 47 out of 235 safety-critical events (SCEs) involved DAS use. An in-depth analysis of 47 SCEs revealed that people misused DASs in 57% of SCEs (e.g., engaged in secondary tasks, used the systems not on highways or with hands off the wheel). During 13% of SCEs, the systems neither reacted to the situation nor warned the driver. A post-study survey showed that the participants found DASs useful and usable. However, the greater the positive attitude toward DAS features, the more participants felt comfortable engaging in secondary tasks. This is a potential unintended side effect of Level 2 DASs given that they still rely on the human driver’s supervision. This study also captured some scenarios where DASs did not meet driver expectations in typical driving situations, such as approaching stopped vehicles and negotiating curves. The findings may inform the development of human-machine interfaces and training programs to reduce the unintended use of DASs and their safety consequences.

Autonomous driving is experiencing a monumental technological shift, making it one of the most highly discussed topics of research. With novel technologies, like advanced driver assistance systems (ADAS), the automotive industry is making significant advancements in both driving comfort and safety. ADAS rely heavily on Frequency Modulated Continuous Waveform (FMCW) radar sensors for speed and range estimation. However, this technology suffers from interference management problem and an inherent ghost target detection with few specific waveforms. This paper introduces a new technique that utilizes Orthogonal Frequency-Division Multiplexing (OFDM) waveform to overcome ghost detection and increase robustness against interference. OFDM leverages maximum length sequence (MLS) (i.e., m-sequence) and provides orthogonality between different transmitted waveforms. Also, incorporating an appropriate scrambled sequence to the proposed waveform further reduces the impact of interference and improves the reliability of target detection. Simulations were conducted to demonstrate proposed algorithm performance for different scenarios, like single- and multiple-neighbors. Furthermore, results were compared and analyzed relative to conventional FMCW in similar scenarios.

Motor vehicle crashes are one of the leading causes of injuries and deaths for police officers. Advanced driver-assistance systems (ADAS) are driving control systems that have been found to improve civilian drivers’ safety; however, the impact of ADAS on police officers’ driving safety has yet to be investigated thoroughly. Disparities between driver states and tasks performed while driving between police and civilian drivers necessitate this distinction. This study identified the types of ADAS used in police vehicles, their impact on officers’ safety, and proposed potential future ADAS features to be implemented in police vehicles. A systematic literature review was conducted using Google Scholar, Compendex, Web of Science, Transport Research International Documentation (TRID), and Google Patents databases to identify the most prevalent police vehicles used in the U.S., available ADAS features in those vehicles, and the impact of ADAS on officers’ safety. A list of recommended ADAS features was developed based on the review of literature, authors’ knowledge and experience in the field, and the findings of an online survey with 73 police officers. Results indicated the addition of multiple ADAS features including the front vehicle detection system, intersection collision avoidance, evasive steering systems, left turn assist, traffic sign detection system, traffic jam assist, two lane and lane-ending detection, wrong-way alert, and autonomous highway driving features have the potential to improve officer safety and performance while driving. However, there was a void of studies focused on ADAS effects on police driving safety which needs to be addressed in future investigations.

Detekcija objekata je ključna tehnologija koja stoji iza naprednih sistema za asistenciju tokom vožnje. U ovom radu je prikazana primena modela konvolucionih neuronskih mreža u rešavanju problema detekcije objekata od interesa iz sekvence slika. Obeležene slike koje sadrže različite scenarije zabeležene tokom dnevne gradske vožnje preuzete su iz CrowdAI baze. Slike su korišćene za učenje modela i ocenu njegovih performansi tokom faze testiranja. Analizirani su rezultati dobijeni korišćenjem detektora sa različitim brojem konvolucionih slojeva i različitim aktivacionim funkcijama neurona u cilju primene ovakvih modela za detekciju objekata učesnika u saobraćaju u realnom vremenu.

This paper presents an approach which combines LiDAR sensors and cameras of a mobile multi-sensor system to obtain information about pedestrians in the vicinity of the sensor platform. Such information can be used, for example, in the context of driver assistance systems. In the first step, our approach starts by using LiDAR sensor data to detect and track pedestrians, benefiting from LiDAR’s capability to directly provide accurate 3D data. After LiDAR-based detection, the approach leverages the typically higher data density provided by 2D cameras to determine the body pose of the detected pedestrians. The approach combines several state-of-the-art machine learning techniques: it uses a neural network and a subsequent voting process to detect pedestrians in LiDAR sensor data. Based on the known geometric constellation of the different sensors and the knowledge of the intrinsic parameters of the cameras, image sections are generated with the respective regions of interest showing only the detected pedestrians. These image sections are then processed with a method for image-based human pose estimation to determine keypoints for different body parts. These keypoints are finally projected from 2D image coordinates to 3D world coordinates using the assignment of the original LiDAR points to a particular pedestrian.

Millimeter-wave (MMW) imaging radar is one of the major environmental sensors in recent self-driving or driver assistance systems and provides environmentally robust sensing even in optically challenging conditions. As a promising three-dimensional (3-D) radar imaging technique, this study focuses on the range points migration (RPM) method, which has advantages in terms of accuracy and low complexity. In the original RPM algorithm, the parameters are manually or empirically determined by considering the sensor configuration and target shape or distance. To address this limitation, we introduced a novel parameter optimization scheme based on the Gaussian mixture model (GMM) and the expectation maximization (EM) algorithm. In addition, we used the ${k}$ -space decomposition-based parameter determination scheme to determine target shape-dependent parameter selection. The results, assuming the human body imaging scenario, showed that our proposed method retains a highly accurate target image without requiring empirical parameter selection.

Automotive radar is used in many applications of advanced driver assistance systems and is considered as one of the key technologies for highly automated driving. An overview of state-of-the-art signal processing in automotive radar is presented along with current research directions and practical challenges. We provide a comprehensive signal model for the multiple-target case using multiple-input multiple-output schemes, and discuss a practical processing chain to calculate the target list. To demonstrate the capabilities of a modern series production high-performance radar sensor, real data examples are given. An overview of conventional target processing and recent research activities in machine learning and deep learning approaches is presented. Additionally, recent methods for practically relevant radar-camera fusion are discussed.

Adaptive Cruise Control (ACC) is one of Advanced Driver Assistance Systems (ADAS) which takes over vehicle longitudinal control under necessary driving scenarios. Vehicle in ACC mode automatically adjusts speed to follow the preceding vehicle based on evaluation of the surrounding traffic. ACC reduces drivers’ workload as well as improves driving safety, energy economy, and traffic flow. This article provides a comprehensive review of the researches on ACC. Firstly, an overview of ACC controller and applied control theories are introduced. Their principles and performances are discussed. Secondly, several application cases of ACC control algorithms are presented. Then validation work including simulation, Hardware-in-the-Loop (HiL) test and on-road experiment is descripted to provide ideas for testing ACC systems for different aims and fidelities. In addition, studies on human-machine interaction are also summarized in this review to provide insights on development of ACC from the perspective of users. At last, challenges and potential directions in this field is discussed, including consideration of vehicle dynamics properties, contradiction between algorithm performance and computation as well as integration of ACC to other intelligent functions on vehicles.

The current paper proposes a data analysis method to evaluate the impact of vehicle-technology evolution on road safety based on a recent and extensive accident dataset (1.3 million vehicles in 201...

Lane detection algorithms play a key role in Advanced Driver Assistance Systems (ADAS), which are however unable to achieve accurate lane recognition in low-light environments. This paper presents a novel deep network structure, namely LLSS-Net (low-light images semantic segmentation), to achieve accurate lane detection in low-light environments. The method integrates a convolutional neural network for low-light image enhancement and a semantic segmentation network for lane detection. The image quality is firstly improved by a low-light image enhancement network and lane features are then extracted using semantic segmentation. Fast lane clustering is finally performed by using the KD tree models. Cityscapes and Tusimple datasets are utilized to demonstrate the robustness of the proposed method. The experimental results show that the proposed method has an excellent performance for lane detection in low-light roads.

The path towards a transport system with fully automated vehicles entails a transitioning period where manual and automated vehicles will coexist in traffic for many years. To realize the many suggested benefits of automated transport, such as lower emissions and improved safety, there is a need for more knowledge concerning the user experience of both manual and automated systems. This study measures and compares driver acceptance of an integrated driver assistance system and a retrofit system based on geofence technology. We focus on two use cases: low-emission zones and school zones. The integrated system was tested by 43 participants driving on a pre-defined test route, using an assisting and visual system with HMI integrated in the dashboard. The retrofit system involved 42 participants and was implemented with an on-board unit with a visual system, using an external screen as HMI. Participants tested the system for eight weeks, including two weeks in black mode. Participants in both trials were experienced drivers and likely to be ‘‘early adopters”, and the acceptance of the systems was evaluated through a survey. Although the research design for the two driver assistance systems was not identical, similar questions were given, making it possible to compare driver acceptance and workload. The analysis shows overall high acceptance among the users for both trials. However, the results show that the integrated system had greater levels of satisfaction, usefulness and usability for low emission zones, as well as greater levels of satisfaction and usefulness for school zones. We also conclude that retrofit systems must improve HMI satisfaction to constitute a viable option for transport authorities when integrating novel ITS technologies.

Automated vehicles are envisioned to be an integral part of the next generation of transportation systems. Whether it is striving for full autonomy or incorporating more advanced driver assistance systems, high-accuracy vehicle localization is essential for automated vehicles to navigate the transportation network safely. In this paper, we propose a reinforcement learning framework to increase GPS localization accuracy. The framework does not make rigid assumptions on the GPS device hardware parameters or motion models, nor does it require infrastructure-based reference locations. The proposed reinforcement learning model learns an optimal strategy to make “corrections” on raw GPS observations. The model uses an efficient confidence-based reward mechanism, which is independent of geolocation, thereby enabling the model to be generalized. We incorporate a map matching-based regularization term to reduce the variance of the reward return. The reinforcement learning model is constructed using the asynchronous advantage actor-critic (A3C) algorithm. A3C provides a parallel training protocol to train the proposed model. The asynchronous reinforcement learning strategy facilitates short training sessions and provides more robust performance. The performance of the proposed model is assessed by comparing it with an extended Kalman filter algorithm as a benchmark model. Our experiments indicate that the proposed reinforcement learning model converges fast, has less prediction variance, and can localize vehicles with 50% less error compared to the benchmark Extended Kalman Filter model.

Driving behavior plays a key role in the interaction between vehicle and driver in transportation systems. Some applications about driving behavior in Advanced Driver Assistance Systems (ADAS) improve driving safety significantly. This paper introduces the driving context and models driving behavior in a combination of cognitive perspective and data-driven perspective. First, we use a cognitive fusion method by adding a delay time module to fuse the environmental information and inside information. To better capture the driving context relationship between outside and inside features, we transfer the behavior prediction task to the sequence labeling task by introducing the visual inertia hypothesis. We propose the Predictive-Bi-LSTM-CRF algorithm which used the Bidirectional Long-Short Term Memory Networks (Bi-LSTM) and Conditional Random Field (CRF) as the loss layer to model the driving behavior. Besides, we define a new comprehensive evaluation metric for the prediction task considering F1-score and the prediction time before maneuver together. Our experiment results achieve the state of art performance on the Brain4Cars dataset and demonstrate the applicability of our theory.

One exciting challenge for future scientists and engineers will be to contribute to the development and introduction of self-driving cars. Under the objective of making young talent attending engineering-focused secondary schools and universities enthusiastic about this highly innovative topic, we have developed an educational platform for advanced driver assistance systems (ADASs) and autonomous driving. Our platform is designed for a miniaturized environment with low cost and scalable complexity, simultaneously providing all of the sensors required to automate a vehicle.

Advanced driver assistance systems should not only make the driving experience safer and more comfortable, it should also have a positive effect on driving behaviors. In this paper, an instructor-like assistance system for collision avoidance is developed and realized on an actual vehicle. The proposed system is activated only if the driver is not operating the vehicle properly when facing a collision risk. The vehicle control is shared by the driver and the assistance system. It is controlled by servomotors. In order to fulfill this requirement, a constraint satisfaction problem (CSP) is proposed and solved based on safe driving constraints and predictive vehicle states. Vehicle motion is predicted by a combination of a dynamics model and a potential field model that reflects the driver’s risk feeling to an obstacle. Improved driving behavior is verified and evaluated quantitatively based on driving simulator data. By comparing the driving data before and after using the assistance system, it is found that distance is increased and speed is reduced when passing an obstacle. As a result, driving behavior becomes safer for collision avoidance due to the system’s instruction. Furthermore, an experiment with an actual vehicle also demonstrates the practicability of the control system and shows the influence of different safe driving constraints.

Traditional surrogate measures of safety (SMoS) cannot fully consider the crash mechanism or fail to reflect the crash probability and crash severity at the same time. In addition, driving risks are constantly changing with driver’s personal driving characteristics and environmental factors. Considering the heterogeneity of drivers, to study the impact of behavioral characteristics and environmental characteristics on the rear-end crash risk is essential to ensure driving safety. In this study, 16,905 car-following events were identified and extracted from Shanghai Naturalistic Driving Study (SH-NDS). A new SMoS, named rear-end crash risk index (RCRI), was then proposed to quantify rear-end crash risk. Based on this measure, a risk comparative analysis was conducted to investigate the impact of factors from different facets in terms of weather, temporal variables, and traffic conditions. Then, a mixed-effects linear regression model was applied to clarify the relationship between rear-end crash risk and its influencing factors. Results show that RCRI can reflect the dynamic changes of rear-end crash risk and can be applied to any car-following scenarios. The comparative analysis indicates that high traffic density, workdays, and morning peaks lead to higher risks. Moreover, results from the mixed-effects linear regression model suggest that driving characteristics, traffic density, day-of-week (workday vs. holiday), and time-of-day (peak hours vs. off-peak hours) had significant effects on driving risks. This study provides a new surrogate safety measure that can better identify rear-end crash risks in a more reliable way and can be applied to real-time crash risk prediction in driver assistance systems. In addition, the results of this study can be used to provide a theoretical basis for the formulation of traffic management strategies to improve driving safety.

Object detection using FMCW (Frequency-modulated continuous wave) radars is of massive importance for the advanced driver assistance systems. However, it is exceptionally challenging due to the diversity of the electromagnetic environment and the existence of the class imbalance in the radar data space. In this paper, we propose a cascaded object detection network to achieve accurate object detection using FMCW radars. Consisting of a ROI generation stage and a final detection stage, the proposed cascaded network can tackle the problem of the class imbalance and detect objects from the range-Doppler or range-velocity space effectively. Besides, we propose a range-velocity regression procedure to improve the performance of the range-velocity localization. Extensive simulation experiments demonstrate that our proposed approach can robustly detect objects from noisy electromagnetic environments with a high localization accuracy.

Traffic sign detection and recognition perform a vital function in real-world driver guidance applications, including driver assistance systems. Research into vision-based traffic sign detection (TSD) and traffic sign recognition (TSR) has gained considerable attention in the scientific community, led mainly by three variables: identification, monitoring, and classification. In addition, TSR provides valuable details and alerts for smart cars including advanced driving assistance (ADAS) and cooperative intelligent transport systems (CITS). Our work will generate high-quality synthetic prohibitory sign images using deep convolutional generative adversarial networks (DCGAN). This paper analyzes and discusses CNN models incorporating different backbone architectures and feature extractors, focusing on Resnet 50 and Densenet for object detection. Assessment of the models provides important information, including mean average accuracy (mAP), workspace capacity, detection period, and the amount of billion floating-point operations (BFLOPS). The maximum average accuracy is 92% (Densenet DCGAN), led by 91% (Resnet 50 DCGAN), 88% (Densenet), and 63% (Resnet 50). We find when using the original image and a synthetic image, accuracy increases, while detection time falls. Our findings show that combining original images and synthetic images in the dataset for training can improve intersection over union (IoU) and traffic sign recognition performance.

Traffic boards and traffic signals are used to maintain proper traffic through busy roads. They help to recognize the rules to follow when driving the vehicle. These signs warn the distracted driver, and prevent his/her actions which could lead to an accident. We have proposed a system which can help recognize these boards and signals at real time thus avoiding major mishap. A real-time automatic sign detection and recognition can help the driver, significantly increasing his/her safety. Lately traffic sign recognition has got an immense interest lately by large scale companies such as Google, Apple and Volkswagen etc. which is driven by the market needs for intelligent applications such as autonomous driving, driver assistance systems (ADAS), mobile mapping, Mobil eye, Apple, etc. Hence, here, we have implemented to do the same with cost efficient manner using Raspberry Pi. The proposed system detects the traffic board or traffic signals, capture its image which through deep learning approach recognizes the same to give result on dashboard as well it gives the measures of distance from front obstacle which helps to implement brake system if obstacle is near. PiCam is used to capture images of traffic sings and is connected to RaspberryPi. Monitor is used to display required output, showing type of sign and distance of collision. This proposal will avoid large number of accidents occurring at bridges and work in progress area due to automated braking system and simultaneous reduce death ratio.

Advanced driver assistance systems (ADAS) promise improved driving performance and safety. With ADAS taking on more vehicle control tasks, the driver’s role may be reduced to that of passive supervision. This in turn may increase drivers’ engagement in non-driving-related tasks, thereby potentially reducing any promised safety benefit. We conducted a systematic review, in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, to study the relationship between ADAS use and driver distraction. Four research questions were addressed—two questions examined the effect of ADAS on secondary task engagement, and the quality of secondary task performance, and two addressed the effects of ADAS on driver attention and on driver behavior changes caused by secondary task engagement. Twenty-nine papers were selected for full text synthesis. The majority of the papers indicate an association between ADAS and increased secondary task engagement, as well as improved secondary task performance. Ten papers reported that drivers tend to divert their attention to secondary tasks and away from driving tasks. These outcomes highlight the continued importance of the role of the human driver despite vehicle automation, especially in the context of driver distraction, and that user understanding of ADAS functionalities and limitations is essential to appropriate and effective use of these systems.

The development of vehicle tracking using sensor fusion is presented in this paper. Advanced driver assistance systems (ADAS) are becoming more popular in recent years. These systems use sensor information for real-time control. To improve the standard and robustness, especially in the presence of environmental noises like varying lighting, weather conditions, and fusion of sensors has been the center of attention in recent studies. Faced with complex traffic conditions, the single sensor has been unable to meet the security requirements of ADAS and autonomous driving. The common environment perception sensors consist of radar, camera, and lidar which have both pros and cons. The sensor fusion is a necessary technology for autonomous driving which provides a better vision and understanding of vehicles surrounding. We mainly focus on highway scenarios that enable an autonomous car to comfortably follow other cars at various speeds while keeping a secure distance and mix the advantages of both sensors with a sensor fusion approach. The radar and vision sensor information are fused to produce robust and accurate measurements. And the experimental results indicate that the comparison of using only radar sensors and sensor fusion of both camera and radar sensors is presented in this paper. The algorithm is described along with simulation results by using MATLAB.