Automated vehicles

The Cyber-Physical Systems (CPSs) and Internet of Things (IoT) have become emerging and essential technologies of the past few decades that connect various heterogeneous systems and devices. Sensors and actuators are fundamental units in most CPS and IoT systems, they are used extensively in vehicle systems, smart health care systems, smart buildings and cities, and many other types of applications. The extensive use of sensors and actuators, coupled with their increasing connectivity, exposes them to a wide range of threats. Given their integration into various systems and the use of multiple technologies, it is very useful to characterize their functions abstractly. For concreteness, we study them here in the context of autonomous cars. An autonomous car is an example of a CPS, which includes IoT applications. For instance, IoT units allow an autonomous car to be connected wirelessly to roadside units, other vehicles, and fog and cloud systems. Also, the IoT allows them to collect and share information on traffic, navigation, roads, and other aspects. An autonomous car is a complex system, not only due to its intricate design but also because it operates in a dynamic environment, interacting with other vehicles and the surrounding infrastructure. To manage these functions, it must integrate various technologies from different sources. Specifically, a diverse array of sensors and actuators is essential for the functionality of autonomous vehicles.

Connectivity and automation have expanded with the development of autonomous vehicle technology. One of several automotive serial protocols that can be used in a wide range of vehicles is the controller area network (CAN). The growing functionality and connectivity of modern vehicles make them more vulnerable to cyberattacks aimed at vehicular networks. The CAN bus protocol is vulnerable to numerous attacks as it lacks security mechanisms by design. It is crucial to design intrusion detection systems (IDS) with high accuracy to detect attacks on the CAN bus. In this dissertation, to address all these concerns, we design an effective machine learning-based IDS scheme for binary classification that utilizes eight supervised ML algorithms, along with ensemble classifiers, to detect normal and abnormal activities in the CAN bus. Moreover, we design an effective ensemble learning-based IDS scheme for detecting and classifying DoS, fuzzing, replay, and spoofing attacks. These are common CAN bus attacks that can threaten the safety of a vehicle’s driver, passengers, and pedestrians. For this purpose, we utilize supervised machine learning in combination with ensemble methods. Ensemble learning aims to achieve better classification results through the use of different classifiers that are combined into a single classifier. Furthermore, in the pursuit of real-time attack detection and classification, we use the Kappa architecture for efficient data processing, enhancing the IDS’s accuracy and effectiveness. We build this system using the most recent CAN intrusion dataset provided by the IEEE DataPort. We carried out the performance evaluation of the proposed system in terms of accuracy, precision, recall, F1-score, and area under curve receiver operator characteristic (ROC-AUC). For the binary classification, the ensemble classifiers outperformed the individual supervised ML classifiers and improved the effectiveness of the classifier. For detecting and classifying CAN bus attacks, the ensemble learning methods resulted in a robust and accurate multiclassification IDS for common CAN bus attacks. The stacking ensemble method outperformed other recently proposed methods, achieving the highest performance. For the real-time attack detection and classification, the ensemble methods significantly enhance the accuracy the real-time CAN bus attack detection and classification. By combining the strengths of multiple models, the stacking ensemble technique outperformed individual supervised models and other ensembles.

Today’s mainstream vehicles are partially automated via an advanced driver assistance feature (ADAS) known as Adaptive Cruise Control (ACC). ACC uses data from on-board sensors to automatically adjust speed to maintain a safe following distance with the preceding vehicle. Contrary to expectations, ICE vehicles equipped with ACC may reduce capacity at bottlenecks because its delayed response and limited initial acceleration during queue discharge could increase the average headway. On the other hand, ACC equipped EVs can potentially mitigate this effect for having ready torque and quicker acceleration. However, this has not been investigated for cases when lane changers enter from the adjacent lane. ACC could respond differently under these conditions, and this car following behavior is often referred as receiving lane change car following. Carefully planned field experiments on lane change car following demonstrate that lane changes and the subsequent receiving lane change car following from ICE vehicles equipped with ACC increases the gap unless the lane changer and the target lane traffic have identical or similar speeds for internal combustion engine (ICE) vehicles and ACC in the EVs doesn’t increase the gap after lane change increasing capacity for merging compared to ICE vehicles. For ICE, this trend also correlates with the selected ACC gap, with larger gap selection resulting in longer gap following the lane change maneuver and the receiving lane change car following in response. Larger gap setting shows better results after lane change for EVs.

Today’s mainstream vehicles are partially automated via an Advanced Driver Assistance Feature (ADAS) known as Adaptive Cruise Control (ACC). ACC relies on data from onboard sensors to automatically adjust speed to maintain a safe following distance with the preceding vehicle. Contrary to expectations for automated vehicles, ACC may reduce capacity at bottlenecks because its delayed response and limited initial acceleration during queue discharge could increase the average headway. Fortunately, when ACC is paired with fully electric vehicles (EVs), EV’s unique powertrain characteristics such as instantaneous torque and aggressive regenerative braking could allow ACC to adopt shorter headways and accelerate more swiftly to maintain shorter headways during queue discharge, therefore reverse the negative impact on capacity. This has been verified in a series of car following field experiments. Field experiments demonstrate that EVs with ACC can achieve a capacity as high as 3333 veh/hr/lane when cruising in steady state conditions at typical freeway speeds (60 mph and 55 mph) and arterial speeds (45 mph and 35 mph). Furthermore, speed fluctuations and disturbances that may come from queues forming at or near the bottleneck do not reduce the capacity, unlike ACC-equipped internal combustion engine (ICE) vehicles, making ACC-equipped EVs outperform ICE vehicles with ACC, as well as human drivers.

Mainstream vehicles sold today are equipped with the Advanced Driver Assistance System (ADAS) known as Adaptive Cruise Control (ACC). ACC automatically adjusts speeds and maintains a safe following distance with the preceding vehicle. This enables partial automation by automating longitudinal car-following. Despite the ever-increasing market penetration, ACC-equipped vehicles will likely operate in a mixed environment with other human-driven vehicles first. However, the traffic flow impact of human driver behavior when following ACC-equipped vehicles is largely unknown, and it is uncertain whether this deserves special consideration when modeling human driver behavior near ACC enabled vehicles. This study conducted a preliminary real-world experiment on a freeway (a portion of Interstate 95) and an urban arterial (a portion of state route A1A) to investigate the human driver behavior with and without the presence of vehicles in ACC mode as the leaders. This unbiased experiment was conducted in naturalistic traffic conditions. Results from the field experiments demonstrate that in a mixed environment with ACC-equipped vehicles as leaders, the human driven vehicles as the follower adopt similar headway, spacing, and acceleration on both freeway and arterial, with no statistically significant difference. The only exception is when traveling at speeds below 15 mph on urban arterials, where human drivers adopt significantly larger spacing while following ACC-enabled vehicles. We expect that findings from these field experiments will provide important initial insights to future research on human driver car following models in a mixed traffic environment and dedicated lanes for automated vehicles.

Adaptive cruise control (ACC) system is the first widely offered automated functionality that regulates the longitudinal movement of the vehicle using onboard radar sensors, and they can maintain a safe following distance with the preceding vehicle. In most of the field experiments with ACC-equipped vehicles conducted with internal combustion engine vehicles, there is still a gap in research on how the automation systems such as ACC combined with electric powertrains will influence the traffic flow be examined.
This study refined and recalibrated an ACC car-following model for EVs and integrated it into AIMSUN to realistically simulate ACC-equipped vehicles and their impact on the fundamental diagram of traffic flow. Simulations were conducted for various ACC market penetrations, and fundamental diagrams were constructed for those market penetrations using detector measurements at various locations along the simulated segment. Overall, the capacity and the jam density increase as the EV with ACC market penetration rises. EVs with ACC can achieve higher capacities compared to ICEs with ACC.

This thesis explores the cross-cultural demands from self-driving cars in regards to their trust, safety, and driving styles. Through the use of international survey data we establish several AI trust and behavior metrics that can be used for understanding cross-cultural expectations from self-driving cars that can potentially address problems of trust between passengers and self-driving cars, social acceptability of self-driving cars, and development of customized autonomous driving technologies. Further this thesis provides a serverless data-collection framework for future research in driving behaviors.

Automated vehicles (AVs) are becoming more common each day as car manufacturers have started to include advanced driving assistant systems (ADAS) in trendline models. The most basic level of vehicle automation includes Adaptive Cruise Control (ACC) can disrupt and change traffic flow. The current study proposes the development of controlled experiments to obtain traffic flow properties for vehicles equipped with ACC in different scenarios. As part of this dissertation, the effects of ACC on capacity are quantified at steady state conditions, meaning cruising speeds or free flow, and at bottlenecks, where speed fluctuations occur. The effects of ACC on traffic flow properties are also assessed by the construction and study of the Fundamental Diagram. Lastly, the vehicles are submitted to less predictable deceleration scenarios that involve a leading vehicle driven in ACC mode and a leading vehicle driven manually. The reaction of ACC for these cases is documented.

The first generation of autonomous vehicles are equipped with Adaptive Cruise Control
(ACC), which automatically adjusts the vehicle speed to maintain a safe following distance and gap selected by the driver. Today’s ACC can also operate at low speeds and signalized intersections on arterial streets. However, the latency of the on-board sensors can significantly increase the start-up lost time and reduce capacity and increase delay on arterials with signalized intersections. This study investigates the fundamental characteristics of traffic flow under ACC vehicles and mixed driving scenarios. Field tests demonstrated that the design of ACC vehicles can lead to delayed response and gradual acceleration when operating on arterials with speed fluctuations due to disturbances. This study also examines the effect of increasing adoption of ACC vehicles at signalized intersections. Field validated simulations suggest that 100% market penetration of ACC vehicles could decrease the capacity by up to 10%. Furthermore, fuel consumption and emissions (CO2, NOx, CO, HC) can increase by up to 33%.

Adaptive Cruise Control (ACC) vehicles have a longer reaction time, and the on-board sensors have a limited detection range that adversely affects the freeway bottleneck capacity. These limitations can cause small speed fluctuations into larger stop-and-go waves at typical freeway bottlenecks. Microsimulation results revealed that flow instability increases with the increase in ACC market penetration for a single lane freeway. The ACC car following model was developed for higher speed ranges only; thus, it could not capture rapid deceleration to lower speeds, let alone complete stops. The algorithm applies collision avoidance and brake relatively late in those instances, which leads to vehicles clustered closer together when at complete stops (or lower speeds). Therefore, the jam density increases with ACC market penetration. Simulation results also represented that no change in capacity was observed with the introduction of ACC vehicles on a freeway without diverging off-ramp and merging on-ramp demand compared to manually driven vehicles. The result is owed to the fact that lane changes and disturbances are not prominent without merging and diverging sections. However, the situation aggravates more for ACC vehicles when there is diverging off-ramp demand and merging on-ramp demand. The effect becomes severe with the increase of ACC market penetration. The field experiments for the fundamental characteristics of traffic flow showed that maximum capacity can be achieved when all the vehicles are operating in ACC mode. However, that maximum flow is unstable, and a minor speed variation can cause severe capacity drop. The jam density is also more in all ACC scenario that might result in rapid queue propagation as the wave speed is larger compared to the mixed driving scenario.