Association Analysis of ADAS and ADS Accidents: A Comparative Study Based on Association Rule Mining

Perception system plays an important role in Advanced driver assistance systems (ADAS) & Autonomous vehicles (AV) to understand the surrounding environment and further navigation. It is highly challenging to achieve the accurate perception of ego vehicle mimicking human vision. The available ADAS and AV solutions could able to perceive the environment to some extent using multiple sensors like Lidars, Radars and Cameras. National Highway Traffic Safety Administration Crash reports of ADAS and AV systems shows that the complete autonomy is challenging to achieve using the existing sensor suite. Particularly, in extreme weather, low light and night scenarios, there is a need for additional perception sensors. Infrared camera seems to be one of the potential sensors to address such extreme and corner cases. This chapter aimed to discuss the advantage of adding infrared sensors to perceive the environment accurately. The advancements in deep learning approaches further leverages to enhance ADAS features. Also, the limitations of current sensors, the need for infrared sensors and technology, artificial intelligence and current research focus using IR images are discussed in detail. Literature shows that by adding IR sensor to existing sensor suite may lead a way to achieve level 3 and above autonomous driving precisely.

ACCIDENT BASED REQUIREMENTS ANALYSIS FOR ADVANCED DRIVER ASSISTANCE SYSTEMS

Analysis of the National Motor Vehicle Crash Causation Survey, conducted by the National Highway Traffic Safety Administration (NHTSA), shows that driver error is a factor in 94% of crashes. Although it is important to remember multiple factors contribute to all crashes, the largest portion of driver error issues involve the driver failing to recognize hazards, including distraction. Around 3,700 people die in traffic every day around the world, and 100,000 are injured. The automotive industry is striving to make driving safer. ADAS in India is comparatively in a nascent stage. However, it is gradually gaining pace. The government's upcoming safety regulations and consumer awareness will give further impetus to this movement. So, Advanced driver-assistance systems (ADAS) is equipping cars and drivers with advance information and technology to make them become aware of the environment and handle potential situations in better way semi-autonomously. High-quality training and test data is essential in the development and validation of ADAS systems which lay the foundation for autonomous driving technology. In addition to this, ADAS systems need to be very safe and robust, with the ability to perform in a variety of driving scenarios, and be very secure, being immune from any external cyber-attacks. In order to make ADAS systems safer, the AV will be required to drive more than a billion miles on real roads, taking tens and sometimes hundreds of years to drive those miles, considering even the most aggressive testing assumptions. Every small update to the AV will require another billion miles of testing to be approved for real world use. Moreover, the more advanced the technology becomes, the more miles will need to de driven. Real word testing plays a very crucial role in ADAS and AV development and testing. Nevertheless, relying only on real world testing will significantly slow down the development and testing of such technologies. This is where simulation comes into play. With the primary objective of road safety improvement, ADAS functionalities will definitely play a big role for automotive industry. In order to tackle Indian specific road infrastructure conditions, and thus improving the safety, a complete tool-chain for developing, deploying and validating ADAS functionalities need to be developed. The presented work shares insights of each and every aspect of this tool-chain with experimental results and real world correlations.

LaneScanNET: A deep-learning approach for simultaneous detection of obstacle-lane states for autonomous driving systems

Automotive industry is a key sector in developed countries, taking advantage from Electronic and Semiconductor industries, for which this work is focused on, including an overview of embedded systems and related technologies for Advanced Driver Assistance Systems (ADAS) development, end user applications and their implementation (SoCs, Application Processors-APs, MCUs, software and boards), manufacturers solutions, architectures, trends and other aspects (like methodologies) to improve functional safety, reliability and performances. The current status to permit the transition from ADAS to Autonomous Driving (AD) systems and Self-Driving Cars (SDC) is also explored.

<div class="section abstract"><div class="htmlview paragraph">Safety is the cornerstone for Advanced Driver Assistance Systems (ADAS) and Autonomous Driving Systems (ADS). To assess the safety of a traffic situation, it is essential to predict motion states of traffic participants in the future with mathematic models. Accurate vehicle trajectory prediction is an important prerequisite for reasonable traffic situation risk assessment and appropriate decision making. Vehicle trajectory prediction methods can be generally divided into motion model based methods and maneuver model based methods. Vehicle trajectory prediction based on motion models can be accurate and reliable only in the short term. While vehicle trajectory prediction based on maneuver models present more satisfactory performance in the long term, these maneuver models rely on machine learning methods. Abundant data should be collected to train the maneuver recognition model, which increases complexity and lowers real-time performance. In this paper, a vehicle trajectory prediction method based on motion model and maneuver model fusion with Interactive Multiple Model (IMM) is proposed. Firstly, Constant Turn Rate and Acceleration (CTRA) motion model and Unscented Kalman Filter (UKF) are used to predict vehicle trajectory with uncertainty in the future. Then, vehicle trajectory prediction based on simplified maneuver recognition model is conducted, using temporal and spatial relationship between vehicle historical trajectory and lane lines. After that, vehicle trajectory prediction by integrating motion model and maneuver model with IMM is conducted. Finally, the proposed method is compared with CTRA motion model based vehicle trajectory prediction and lane keeping model (LKM) based vehicle trajectory prediction in two simulation test scenarios. The simulation results indicates that the IMM-based method achieves both excellent prediction accuracy and appropriate prediction uncertainty in the whole prediction horizon. This research can be used to support decision making for Advanced Driver Assistance Systems (ADAS) and Autonomous Driving Systems and leads to improvement of traffic safety.</div></div>

Traffic signs are an essential component of road infrastructure. According to the Department of Transportation, Indonesia has over 300 distinct traffic signs, categorized based on their functions and purposes. TSR systems have been widely integrated into various intelligent transportation technologies, such as Driver Assistance Systems (DAS), Advanced Driver Assistance Systems (ADAS), and Autonomous Driving Systems (ADS). The output generated by TSR serves as a critical input for DAS, ADAS, ADS, and other intelligent systems. This article presents a CNN-based classification for traffic sign recognition using multi-task learning (MTL), focusing on traffic signs in Indonesia. The dataset was collected from direct capture with the help of a cellphone camera, indirect capture by utilizing screenshots on a digital map application, and they are captured from several different angles, during the day and at night. The proposed CNN architecture incorporates multi-scale within an MTL framework. The use of a multi-scale approach will hopefully enhance the model’s ability to recognize traffic signs in varied and complex environments. And the integration of MTL will enable the model to handle multiple related tasks concurrently, sharing learned features across tasks. During the training stage, the MS-CNN outperformed a standard CNN model by demonstrating lower initial loss, higher starting accuracy, and achieving 100% accuracy by the 8th epoch with a minimal error rate of just 0.003. In the testing stage, the model achieved exceptional results, as shown by the confusion matrix, it successfully classified all traffic sign types (10 classes) and accurately categorized each sign into one of two categories—warning or prohibition. All performance metrics, including precision, recall, and F1-score, reached 100% for both output tasks, confirming the robustness and reliability of the model.

<div class="section abstract"><div class="htmlview paragraph">While various Advanced Driver Assistance System (ADAS) features have become more prevalent in passenger vehicles, their ability to potentially avoid or mitigate vehicle crashes has limitations. Due to current technological limitations, forward collision mitigation technologies such as Forward Collision Warning (FCW) and Automated Emergency Braking (AEB) lack the ability to consistently perform in many unique and challenging scenarios. These limitations are often outlined in driver manuals for ADAS equipped vehicles. One such scenario is the case of a stationary lead vehicle at the side of the road. This is generally considered to be a challenging scenario for FCW and AEB to address because it can often be difficult for the system to discern this threat accurately and consistently from non-threatening roadway infrastructure without unnecessary or nuisance system activations. This is made more difficult when the stationary lead vehicle is only partially in the driving lane and not directly in the forward path, as is the case in the current FCW and AEB confirmation test protocols used by the National Highway Traffic Safety Administration (NHTSA). Partial overlap tests are carried out by Euro New Car Assessment Program (NCAP), however data has not been published to date showing the effect this has on performance.</div><div class="htmlview paragraph">A test series was designed to investigate the effect of variable overlap, with stationary lead vehicle targets, on the triggering and timing of warnings presented by forward collision mitigation technology.</div></div>

With the increasing number of vehicles equipped with automated driving systems (ADS) being tested on public roads and the expanding market share of vehicles equipped with advanced driving assistance systems (ADAS), the number of ADS- or ADAS-involved crashes increases. Thus, it is necessary to investigate the distribution of ADS- and ADAS-involved crashes and the factors leading to them. The rear-end collision has been found to dominate ADS-involved crashes. However, no research has explored the conditions when ADS-involved rear-end collisions are more likely to happen and no research has investigated ADAS-involved rear-end crashes. Based on 130 ADS-involved crashes and 84 ADAS-involved crashes extracted from a dataset collected by the National Highway Traffic Safety Administration (NHTSA) between July 2021 and May 2022, this study explored the crash patterns, especially rear-end crashes, of ADS- and ADAS-controlled vehicles. Results show that rear-end collisions dominate both ADS- and ADAS-involved crashes, especially ADAS-involved crashes. The type of both ADS-involved and ADAS-involved crashes was affected by the speed of the ego-vehicle relative to the posted speed limit. Further, the type of ADS-involved crash was affected by the pre-crash movement of the crash partner, while the type of ADAS-involved crash was further associated with the road type. The findings can provide insights into the design of ADAS and ADS control algorithms, the external human-machine interface design of the vehicles with ADS or ADAS, and the training program of human road users to improve traffic safety in mixed traffic.

To support safe driving, numerous methods have been proposed for detecting distractions based on the measure- ments of a driver's gaze. These methods empirically focused on certain driving contexts, and analyzed gaze behavior under particular peripheral vehicular conditions; therefore, numerous driving situations were not considered. To address this problem, we propose a data-driven approach that analyzes peripheral vehicular behaviors during gaze transitions of drivers, to compare their neutral driving state with a cognitive distraction state. The analysis results show that drivers, under the neutral conditions, turned their gaze to peripheral vehicles to be focused on; however, they did not do this consistently under the distracted conditions. In addition, we propose a simple classifier to discriminate between the distracted and the neutral states by analyzing peripheral vehicular behavior. The proposed classifier can manage various situations, and provide high discrimination accuracy, by focusing on gaze transitions from the front view toward other directions. I. INTRODUCTION Advanced Driver Assistance Systems (ADAS) and autopi- lot vehicles have attracted significant attention. However, researchers are concerned that drivers may develop an over- reliance on incomplete systems, which may lead to accidents. To prevent overreliance, ADAS must effectively analyze the driver's state as well as traffic situations. In this study, we focus on driver's cognitive distraction, which is an observable state resulting from the overreliance, and its relationships with driver's eye-gaze and peripheral vehicular behaviors. Driver distraction is a diversion of attention away from activities critical for safe driving toward a competing activity(1), and is a significant risk factor that can cause accidents(2). Note that distraction differs from fatigue(3), which is defined as a state of exhaustion that disables a person from continuing an activity(4). Numerous researchers have developed driver distraction monitoring systems that aim to promote driving safety by considering different types and levels of distraction(3). The National Highway Traffic Safety Administration (NHTSA) classifies distractions into the following categories: (1) visual distraction; (2) auditory distraction; (3) biomechanical distraction; and (4) cognitive distraction from the viewpoint of the driver's functionality(2). Visual distraction, auditory distraction, and biomechanical distraction are caused by external factors that disturb the

As operators of motor vehicles, drivers have been described as “outdated . . . with Stone Age characteristics and performance . . . controlling a fast, heavy machine in an environment packed with unnatural, artificial signs and signals” (Rumar, 1981). Despite our anatomical, physiological, and perceptual shortfalls, the fatality rate in the United States hit a historic low of 1.1 fatalities per 100 million vehicle miles traveled (VMT) in 2011 (National Highway Traffic Safety Administration, 2011). Fatal crash involvement by VMT increases by age starting in the mid-60s (Insurance Institute for Highway Safety, 2011), and many individuals begin to curtail or stop driving. However, the cessation of driving due to advanced age comes at great cost. Agerelated losses in the ability to drive equates, at best, to a forfeiture of personal freedom, reliance on the assistance of others to meet basic activities of daily living, and can lead to increased symptoms of depression (Marottoli et al., 1997). Transitions in driving roles occur throughout one’s lifetime. As medical conditions accrue, they can sporadically or permanently limit driving (Owsley, 2004). Women frequently cease driving earlier than men, and often while still fit to drive (Alsnih & Hensher, 2003; Siren, HakamiesBlomqvist, & Lindeman, 2004). Widowhood can increase older women’s need to drive (Braitman & Williams, 2011) at a time when this is particularly challenging. On the other hand, even as adults age, they are becoming increasingly economically able to purchase new vehicles (Coughlin, 2009). As a consequence of both the increased numbers and economic independence of older adults, innovations in personal mobility that mitigate the burdens of age will grow in value over the coming decades. A move toward new urbanism, including improved public transit systems and walkable streets and sidewalks, is an admirable vision that would help meet the growing needs of many older adults. However, it will require, at considerable cost, rebuilding or retrofitting the existing infrastructure at a rate that is not likely to meet the needs of today’s aging boomers. Fully automated or driverless cars, by contrast, represent a path that promises to enhance the mobility options of older adults within the existing infrastructure. However, many consumers do not clearly understand that while the basic building blocks of these systems are available today in advanced driver assistance systems (ADAS), fully automated or driverless vehicles are still on the distant horizon. For the foreseeable future, automated vehicle technologies, including ADAS, will continue to rely on a “responsible” driver to oversee the technology, capable of resuming control and having the foresight to make many (yet to be defined) strategic operational decisions. But because of their transformative promise and heavy news coverage, the prospect of automated cars has become a source of great hope for many. Some believe that fully automated cars, capable of navigating the roadways while the “operator” reads a paper or takes a nap, will be available within a few years. Unfortunately, that is not the case. Instead, there is work to be done to increase the awareness and education necessary to spur the purchasing of ADAS available today, which will support many older drivers’ mobility and safety needs.

<div class="section abstract"><div class="htmlview paragraph">Advanced driver assistance systems (ADAS) and automated driving systems (ADS) continue to expand into the market at a rapid pace. As improved (i.e., next generation) versions of these systems become available, they will continue to face many challenges in their implementation and benefits for safety and driving operations. The solution will involve many parties, including road safety professionals and researchers who see the potential in these systems but may have difficulties keeping up with them, and safety advocates who are calling for these systems to achieve higher levels of safety now.</div><div class="htmlview paragraph"><b>The Challenges of Next-gen ADAS and ADS and Related Vehicle Safety Topics</b> explores these challenges that will fall on the National Highway Traffic Safety Administration (NHTSA) and automakers as they balance costs and benefits; establish reasonable regulations and standards; and determine how to improve, test, deliver, and use these systems successfully. Perhaps the most formidable challenge will be overcoming the expectation that ADAS—and especially ADS—will always work perfectly in every scenario.</div><div class="htmlview paragraph"><a href="https://www.sae.org/publications/edge-research-reports" target="_blank">Click here to access the full SAE EDGE</a><sup>TM</sup><a href="https://www.sae.org/publications/edge-research-reports" target="_blank"> Research Report portfolio.</a></div></div>

Surveys [8] show that people generally have a positive attitude towards autonomous cars. However, these studies neglect that cars have different levels of autonomy and that User Acceptance (UA) and User Experience (UX) with autonomous systems differ with regard to the degree of system autonomy. The National Highway Traffic Safety Administration (NHTSA) defines five degrees of car autonomy which vary in the penetration of cars with Advanced Driver Assistance Systems (ADAS) and the extent to which a car is taken over by autonomous systems. Based on these levels, we conducted an online-questionnaire study (N = 336), in which we investigated how UA and UX factors, such as Perceived Ease of Use, Attitude Towards using the system, Perceived Behavioral Control, Behavioral Intention to use a system, Trust and Fun, differ with regard to the degree of autonomy in cars. We show that UA and UX are highest in levels of autonomy that already have been deployed in modern cars. More specifically, perceived control and fun decrease continuously with higher autonomy. Furthermore, our results indicate that pre-experience with ADAS and demographics, such as age and gender, have an influence on UA and UX.

Event Abstract Back to Event Predicting response latency using EEG alpha-band power and low-cost wearable physiological sensors. Dean Cisler1*, Pamela M. Greenwood1, Ryan McKendrick2 and Carryl L. Baldwin1 1 George Mason University, United States 2 Northrop Grumman (United States), United States According to the National Highway Traffic Safety Administration (NHTSA), over 37,000 people died in traffic crashes in 2016 (NHTSA, 2017). Human error plays a role in a substantial portion of these fatalities with some estimates as high as 94% (Singh, 2015). Advanced driver assistance systems (ADASs, e.g., adaptive cruise control and automatic emergency braking) have the potential to improve driver performance and occupant safety. Automatic emergency braking has been found to reduce rear-end crashes by about 40% (Cicchino, 2017). Although ADASs do reduce driver workload, drivers may consequently become complacent and inattentive. Automated systems are not infallible and over-trust in those systems could lead the driver to fail to monitor or detect critical signals related to system reliability (Parasuraman & Manzey, 2010). Inattention in a partially autonomous vehicle can be hazardous, as current ADASs are not designed to brake effectively during “cut-in,” “cut-out,” and crossing-path scenarios. Now that most new vehicles are equipped with some ADASs, it is important to understand how drivers respond to signals of automation unreliability. Inattentive drivers may require more urgent warnings – warnings that could annoy or startle the attentive driver. One approach to monitoring driver attentional state is to use non-invasive physiological measures. Past research used EEG, eye-tracking, and heart rate variability (HRV) to classify driver states (Hogervorst, Brouwer & van Erp, 2014). Physiological measures can determine whether a driver is on-task or mind wandering (Baldwin, et al., 2017) and can successfully adapt automation to improve performance in an unmanned aerial vehicle task (Wilson & Russell, 2003a, b). EEG alpha-band power can predict driver errors (O’Connell et al., 2009) and predict mind wandering (Baldwin, et al., 2017). However, EEG recording is impractical for real-world driving. Driver attentional state could be monitored with low-cost physiological metrics of gaze dispersion and heart rate variability, previously found to predict speed of “take-over” from automation (Dehais, Causse, & Tremblay, 2011). The current study tested the hypothesis that HRV and gaze dispersion measured with wearable technology (ZephyrTM BioPatch HP Monitoring Device and Pupil Headset by Pupil Labs, respectively) would be as effective as EEG alpha-band power in predicting performance during a simulated fully autonomous lane-change task. We used backward stepwise multiple regression to test the hypothesis that the wearable technology metrics (eye gaze, HRV) would predict the performance data as well as the EEG measure of alpha-band power. Task. The participants operated a driving simulator for five autonomous drives (roughly 11 minutes each). During the drives, an automation interface was presented in the lower right of the windshield composed of right or left facing arrows (170 ms) with varying amounts of red and green. The automation “reliable” arrow tip was both red and green while the automation “unreliable” arrow tip was all red. The participant’s task during the drives was to identify cues indicating potential automation failures (i.e., the unreliable arrows) by pressing a button on the steering wheel when automation unreliable arrows were detected. Reliable automation arrows were presented on 90% of the trials while unreliable arrows were presented on the remaining 10% of trials. Following unreliable arrows, 60% of the time the vehicle did not change lanes, 20% of the time the vehicle made an incorrect lane change, 20% of the time the vehicle made a correct lane change. The participant was asked to make serial button presses to indicate (1) automation unreliability and (2) the error the vehicle made. Before the experiment, participants were equipped with an eye-tracker, heart-rate monitor, and connected to a 40-channel NuAmps EEG system (Fz, Cz, Pz, Oz, P3, P4, F3, F4, R and L mastoids). In addition to responding to the interface arrows, participants were required to (a) keep a running count of billboards and (b) answer situation awareness questions presented on the screen. Response time (RT) to the unreliable arrows, alpha-band at Pz, vertical and horizontal gaze dispersion HRV, and billboard and situation awareness responses were assessed. HRV was calculated as the root mean square of successive differences 10s prior to each unreliable stimulus. Horizontal and vertical gaze dispersion were calculated by taking the log transform of the standard deviation of eye movements 10s before the stimuli (lnX and lnY). Results. Data were analyzed using SPSS. Insofar as performance accuracy was at ceiling for each participant, RT was the behavioral measure of interest. To compare the physiological metrics, we used a backwards regression predicting RT from Cz Alpha and Pz Alpha, HRV, lnY and lnX. This analysis showed that HRV explained 11.3% of the variance (R2 = .113, F(1, 25) = 4.321, p < 0.05) and was the only factor to significantly predict RT (β = 0.384, p < 0.05). Discussion. These results indicate wearable, low fidelity technology show promise for predicting the speed with which a driver in an autonomous vehicle will respond to signals of automation failure. Future applications may be able to use low cost wearables capable of calculating HRV in real time to signal periods of driver inattention. Further work is needed to determine the most efficient method of reorienting drivers’ attention during automation failures. Keywords: alpha-band, Wearable Technology, Autonomous Driving, response latency, Predicting Behavior Conference: 2nd International Neuroergonomics Conference, Philadelphia, PA, United States, 27 Jun - 29 Jun, 2018. Presentation Type: Poster Presentation Topic: Neuroergonomics Citation: Cisler D, Greenwood PM, McKendrick R and Baldwin CL (2019). Predicting response latency using EEG alpha-band power and low-cost wearable physiological sensors.. Conference Abstract: 2nd International Neuroergonomics Conference. doi: 10.3389/conf.fnhum.2018.227.00056 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 11 Apr 2018; Published Online: 27 Sep 2019. * Correspondence: Mr. Dean Cisler, George Mason University, Fairfax, United States, dcisler@masonlive.gmu.edu Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Dean Cisler Pamela M Greenwood Ryan McKendrick Carryl L Baldwin Google Dean Cisler Pamela M Greenwood Ryan McKendrick Carryl L Baldwin Google Scholar Dean Cisler Pamela M Greenwood Ryan McKendrick Carryl L Baldwin PubMed Dean Cisler Pamela M Greenwood Ryan McKendrick Carryl L Baldwin Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

This paper describes recent progress toward achieving representative and reliable active safety performance assessment of advanced driver assistance systems (ADAS). Because ADAS act within a complex, dynamic traffic environment, reliable evaluation of their safety benefits poses methodological challenges. For a proposed ADAS, its expected contribution to reduction of mortality and injuries as well as false positives should be predicted. To meet these challenges, our approach incorporates identification of target scenarios; calibration and validation of stochastic behavior and accident injury models; stochastic (Monte-Carlo) simulation of target scenarios in varied traffic contexts with/without ADAS; and integration of supporting and corroborating field and laboratory analyses. These include a new controlled, high-throughput approach to sensor testing and algorithm validation in camera-based ADAS using a virtual graphical test bed, which supports systematic identification of critical external conditions that could modify performance or lead to a failure mode. The methodologies introduced here are designed to ensure validity of all key links in the assessment chain, not limited to those aspects that can be assessed in a single test.