SmartDriveNet: an integrated approach for robust driving perception in unstructured environments and adverse weather conditions

Urban green spaces are an important part of the urban ecosystem and hold significant ecological value. To effectively protect these green spaces, urban managers urgently need to identify them and monitor their changes. Common urban vegetation positioning methods use deep learning segmentation models to process street view data in urban areas, but this is usually inefficient and inaccurate. The main reason is that they are not applicable to the variable climate of urban scenarios, especially performing poorly in adverse weather conditions such as heavy fog that are common in cities. Additionally, these algorithms also have performance limitations such as inaccurate boundary area positioning. To address these challenges, we propose the UGSAM method that utilizes the high-performance multimodal large language model, the Segment Anything Model (i.e., SAM). In the UGSAM, a dual-branch defogging network WRPM is incorporated, which consists of the dense fog network FFA-Net, the light fog network LS-UNet, and the feature fusion network FIM, achieving precise identification of vegetation areas in adverse urban weather conditions. Moreover, we have designed a micro-correction network SCP-Net suitable for specific urban scenarios to further improve the accuracy of urban vegetation positioning. The UGSAM was compared with three classic deep learning algorithms and the SAM. Experimental results show that under adverse weather conditions, the UGSAM performs best in OA (0.8615), mIoU (0.8490), recall (0.9345), and precision (0.9027), surpassing the baseline model FCN (OA improvement 28.19%) and PointNet++ (OA improvement 30.02%). Compared with the SAM, the UGSAM improves the segmentation accuracy by 16.29% under adverse weather conditions and by 1.03% under good weather conditions. This method is expected to play a key role in the analysis of urban green spaces under adverse weather conditions and provide innovative insights for urban development.

BackgroundMotor vehicle crashes are a leading cause of injury mortality. Adverse weather and road conditions have the potential to affect the likelihood of motor vehicle fatalities through several pathways. However, there remains a dearth of assessments associating adverse weather conditions to fatal crashes in the United States. We assessed trends in motor vehicle fatalities associated with adverse weather and present spatial variation in fatality rates by state.MethodsWe analyzed the Fatality Analysis Reporting System (FARS) datasets from 1994 to 2012 produced by the National Highway Traffic Safety Administration (NHTSA) that contains reported weather information for each fatal crash. For each year, we estimated the fatal crashes that were associated with adverse weather conditions. We stratified these fatalities by months to examine seasonal patterns. We calculated state-specific rates using annual vehicle miles traveled data for all fatalities and for those related to adverse weather to examine spatial variations in fatality rates. To investigate the role of adverse weather as an independent risk factor for fatal crashes, we calculated odds ratios for known risk factors (e.g., alcohol and drug use, no restraint use, poor driving records, poor light conditions, highway driving) to be reported along with adverse weather.ResultsTotal and adverse weather-related fatalities decreased over 1994–2012. Adverse weather-related fatalities constituted about 16 % of total fatalities on average over the study period. On average, 65 % of adverse weather-related fatalities happened between November and April, with rain/wet conditions more frequently reported than snow/icy conditions. The spatial distribution of fatalities associated with adverse weather by state was different than the distribution of total fatalities. Involvement of alcohol or drugs, no restraint use, and speeding were less likely to co-occur with fatalities during adverse weather conditions.ConclusionsWhile adverse weather is reported for a large number of motor vehicle fatalities for the US, the type of adverse weather and the rate of associated fatality vary geographically. These fatalities may be addressed and potentially prevented by modifying speed limits during inclement weather, improving road surfacing, ice and snow removal, and providing transit alternatives, but the impact of potential interventions requires further research.Electronic supplementary materialThe online version of this article (doi:10.1186/s12940-016-0189-x) contains supplementary material, which is available to authorized users.

The International Maritime Organisation (IMO) requirements for the control of greenhouse gas (GHG) emissions of shipping have raised interest in ship manoeuvrability in adverse weather conditions when compliance is accomplished simply by reducing the main engine power. In response, the IMO has adopted the guidelines for determining minimum propulsion power to maintain the manoeuvrability of ships in adverse conditions. In the present paper, a systematic investigation on the manoeuvrability of a ship with different low advance speeds in adverse weather conditions was conducted by means of an unsteady Reynolds-Averaged Navier-Stokes solver. The numerical results demonstrated the contribution of low advance speeds to the course-keeping and turning circle manoeuvre, providing a practical insight into the manoeuvring performance of a ship with minimum propulsion power in adverse weather conditions. For the course-keeping control, the ship experienced more aggressive steering as the propeller revolution decreased in the oblique waves, while it appeared that the difference in the rudder deflection according to the change in the propeller speed in the head, beam, and following waves is negligible. The difficulty of the low speed turning manoeuvre was clearly noted when the direction of the incident wave was opposite to the direction towards which the ship intended to turn. It is believed that this paper can also be impactful in improving the guidelines of minimum powering of ships for safe navigation in adverse weather conditions.

The impact of weather on summer and winter exercise behaviors

Path-following control problem for maritime autonomous surface ships (MASS) in adverse weather conditions at low speeds

Identifying crash-prone traffic conditions under different weather on freeways

Localization-aware logit mimicking for object detection in adverse weather conditions

ABSTRACTWyoming's Interstate 80 has one of the highest truck crash rates in the United States. This is due to a variety of reasons, including high percentage of truck traffic, adverse weather conditions and mountainous terrain. These factors have caused the Wyoming Highway Patrol (WHP) to spend extensive resources on inspecting commercial vehicles and enforcement of traffic laws in this corridor. This study estimated the correlation between traffic citations and truck crashes. In addition, the paper evaluated the increased risk of truck crashes in adverse weather and road conditions. The explanatory variables included geometric features, weather condition, traffic volume and types of citations. This research concluded that speeding related citations and truck crashes are negatively correlated, and the risk of truck crashes is significantly higher when weather is not clear, and the road is not dry.

The degree of autonomy in vehicles depends directly on the performance of their sensor systems. The transition to even more autonomously driven cars therefore requires the development of robust sensor systems with different skills. Especially in adverse and changing weather conditions (rain, snow, fog, etc.), conventional sensor systems such as cameras perform unreliably. Moreover, data evaluation has to be performed in real-time, i.e. within a fraction of seconds, in order to safely guide the car through traffic and to avoid a crash with any obstacle. Therefore, we propose to use a so called time-gated-single-pixel-camera, which combines the principles of time gating and compressed sensing. In a single pixel camera, the amount of recorded data can be significantly reduced compared to a conventional camera by exploiting the inherent sparsity of scenes. The lateral information is gained with the help of binary masks in front of a simple photodiode. We optimize the pattern of the masks by including them as trainable parameters within our data evaluation neural network. Additionally, our camera is able to cope with adverse weather conditions due to the underlying time gating principle. The feasibility of our method is demonstrated by simulated and measured data as well.

The impact of adverse weather conditions on transportation operation and safety is the focus of many studies; however, comprehensive research detailing the differences in driving behavior and performance during adverse conditions is limited. Many previous studies utilized aggregate traffic and weather data (e.g., average speed, headway, and global weather information) to formulate conclusions about the impact of weather on network operation and safety; however, research into specific factors associated with driver performance and behavior are notably absent. A novel approach, presented in this paper, fills this gap by considering disaggregate trajectory-level data available through the SHRP2 Naturalistic Driving Study and Roadway Information Database. Parametric ordinal logistic regression and non-parametric classification tree modeling were utilized to better understand speed selection behavior in adverse weather conditions. The results indicate that the most important factors impacting driver speed selection are weather conditions, traffic conditions, and the posted speed limit. Moreover, it was found that drivers are more likely to significantly reduce their speed in snowy weather conditions, as compared with other adverse weather conditions (such as rain and fog). The purpose of this study was to gather insights into driver speed preferences in different weather conditions, such that efficient logic can be introduced for a realistic variable speed limit system—aimed at maximizing speed compliance and reducing speed variations. This study provides valuable information related to drivers’ interaction with real-time changes in roadway and weather conditions, leading to a better understanding of the effectiveness of operational countermeasures.

Driving is important for older people to maintain mobility. To reduce age‐related functional decline, older drivers may adjust their driving by avoiding difficult situations. One of these situations is driving in adverse weather conditions such as in the rain, snow and fog which reduce the visual clarity of the road ahead. The upcoming highly automated vehicle (HAV) has the potential of supporting older people. However, only limited work has been done to study older drivers’ interaction with HAV, especially in adverse weather conditions. This study investigates the effect of age and weather on takeover control performance among drivers from HAV. A driving simulation study with 76 drivers has been implemented. The participants took over the vehicle control from HAV under four weather conditions clear weather, rain, snow and fog, where the time and quality of the takeover control are quantified and measured. Results show age did affect the takeover time (TOT) and quality. Moreover, adverse weather conditions, especially snow and fog, lead to a longer TOT and worst takeover quality. The results highlighted that a user‐centred design of human–machine interaction would have the potential to facilitate a safe interaction with HAV under the adverse weather for older drivers.

<div class="section abstract"><div class="htmlview paragraph">Studies of automatic emergency braking (AEB) find that AEB-equipped vehicles are around half as likely to crash. This study examines whether driver characteristics and road and weather conditions modify this preventive effect of AEB.</div><div class="htmlview paragraph">Toyota production data were merged with police reported crash files from eight U.S. states for crash years 2015 up to 2019 by 17-digit vehicle identification number (VIN). Using a case-control design, this study investigated the relationship of AEB presence with being a case vehicle in a system-relevant crash (the striking vehicle in front-to-rear crash; n=30,056) versus an AEB non-relevant control vehicle (the struck vehicle in a front-to-rear crash; n=62,820). The analysis was stratified by driver characteristics and by weather and road conditions. Logistic regression modeled the relationship, controlling for exposure (vehicle-days) and possible confounding factors. The resulting odds ratios for AEB equipment from the separate models were compared to determine if the effect of AEB presence was modified by the characteristic or condition of interest.</div><div class="htmlview paragraph">Overall, AEB-equipped vehicles were 43% (p<0.001) less likely to be the striking (case) vehicle compared to non-equipped vehicles. However, the preventive effect of AEB was significantly lower among older drivers (over 65 years) compared to younger drivers; 29% less likely to be a striking vehicle (OR=0.71) versus 46% (OR=0.54), respectively. The effect of AEB was also lower in adverse weather conditions (rain, fog, snow) (OR=0.66) and on wet or snowy roads (OR=0.65), though these differences were not significant compared to clear weather and dry roads. The AEB effect was also lower among risk-taking drivers (alcohol-involved, speeding, or unrestrained) compared to non-risk-taking (OR=0.72 versus OR=0.59, respectively).</div><div class="htmlview paragraph">AEB prevents crashes, regardless of driver characteristics and environmental conditions. This study suggests, however, that the size of the effect is smaller among older and risk-taking drivers, and in adverse weather and road conditions.</div></div>

Automated anomaly detection in surveillance videos has attracted much interest as it provides a scalable alternative to manual monitoring. Most existing approaches achieve good performance on clean benchmark datasets recorded in well-controlled environments. However, detecting anomalies is much more challenging in the real world. Adverse weather conditions like rain or changing brightness levels cause a significant shift in the input data distribution, which in turn can lead to the detector model incorrectly reporting high anomaly scores. Additionally, surveillance cameras are usually deployed in evolving environments such as a city street of which the appearance changes over time because of seasonal changes or roadworks. The anomaly detection model will need to be updated periodically to deal with these issues. In this paper, we introduce a multi-branch model that is equipped with a trainable preprocessing step and multiple identical branches for detecting anomalies during day and night as well as in sunny and rainy conditions. We experimentally validate our approach on a distorted version of the Avenue dataset and provide qualitative results on real-world surveillance camera data. Experimental results show that our method outperforms the existing methods in terms of detection accuracy while being faster and more robust on scenes with varying visibility.

In response to extreme traffic congestion in metropolitan areas that causes unnecessarily long travel times, high fuel consumption, and excessive greenhouse gas emissions, transportation agencies have implemented various strategies to mitigate traffic congestion. Managed lanes—one of the measures applied worldwide—provide benefits to road users and operating agencies by integrating advanced technologies such as electronic and dynamic tolling systems. However, those agencies already implementing or considering implementing the managed lane strategy are seeking a solution to effectively and properly charge toll rates based on vehicle occupancy and penalize violating vehicles. Vehicle passenger detection systems (VPDSs) have been developed and evaluated worldwide, but limitations still inhibit their full implementation. This study confirms that the performance of the deep learning algorithm, a core VPDS technology, declines under certain adverse weather conditions because of lack of training data sets. The performance of the “you only look once” (YOLOv3) model trained with a normal weather data set decreased by as much as 8.5% when it was tested for adverse weather conditions. In this study, augmented reality (AR) models are developed to enhance the accuracy of vehicle passenger detection (VPDA) by the VPDS by training the algorithm with AR images representing virtual adverse weather conditions. Models trained with AR image sets of various weather categories (fog, rain, and snow) attained VPDA enhanced by up to 7.9%. The final model significantly improves VPDA under adverse weather conditions. The proposed models could be considered for implementation with road weather information systems under adverse weather conditions.

Exploring factors contributing to injury severity at work zones considering adverse weather conditions