Human-driven Vehicles Research Articles

Real-time executable platoon formation approach using hierarchical cooperative motion planning framework

While connected and automated vehicle (CAV) platooning holds promise for enhancing traffic efficiency and reducing energy consumption, we still lack efficient algorithms for guiding the local movements of CAVs to form a platoon on a road due to significant computational and control challenges. This study addresses this gap by designing a real-time executable Hierarchical and Recursive Platoon Formation (HR-PF) framework tailored to mixed flow traffic conditions that encompass both Human-Driven Vehicles (HDV) and CAVs. The HR-PF framework comprises three hierarchical mathematical models (modules) designed to optimize platoon formation while considering both macroscopic traffic conditions and microscopic traffic safety. Module-I formulates a mixed integer quadratic program to determine the timing, location, and state of platoon formation. It is further extended to a mixed integer nonlinear program so that we can also select the optimal size of the target platoon. Module-II designs a Hybrid State-Lattice Motion planner to generate optimal trajectory references for CAVs to approach the target platoon state, ensuring microscopic traffic safety. Module-III develops longitudinal and lateral controllers to enable CAVs to track trajectory references accurately. These models function recursively at varying frequencies to balance mathematical rigor with practical application. Numerical experiments demonstrate that HR-PF facilitates efficient platoon formation in real-time across diverse traffic scenarios and road geometries while sustaining traffic efficiency. Furthermore, the performance of platoon formation is affected by surrounding traffic density and CAV penetrations, with prompt formation observed under LOS C and D traffic environments and slightly more traffic impacts under LOS E and F. These findings provide robust support for exploring advanced platoon formation and platooning strategies for CAVs under complicated traffic environments.

FAIRLANE: A multi-agent approach to priority lane management in diverse traffic composition

The rise of autonomous driving technologies prompts a reevaluation of traditional urban traffic control and lane management. Dedicated lanes for connected and autonomous vehicles (CAVs), with intermittent access for other vehicles, have been proven to enhance road capacity and reduce underutilization moderately. However, this assumes all non-CAVs are smart vehicles, which is different from the current baseline for the street vehicle mix. Presently, our streets feature a mix of CAVs, smart vehicles, and human-driven vehicles, and the research on dedicated lanes using the realistic mixed traffic environment is missing. In this paper, we investigate the enhancement of road utilization using realistic mixed traffic combinations and identify the penetration rate of CAVs and smart vehicles necessary to improve baseline utilization. Previous studies have focused on lane-management strategies in single-vehicle settings, neglecting the interaction of CAVs with neighboring CAVs and smart vehicles. Therefore, we propose a multi-agent reinforcement learning-based framework to facilitate fair utilization of priority lanes, considering driving comfort, traffic efficiency, and safety during lane-changing. Through multiple experiments on a realistic network, our results demonstrate that the proposed framework significantly improves traffic efficiency, particularly when the penetration rate of CAVs is below 40% and Semi-Autonomous Vehicles (SAVs) constitute 50% of the remaining vehicles. The framework outperforms traditional lane management strategies, reducing mean waiting time and increasing average speed. This study provides nuanced information on different vehicle penetration rates, enabling more informed decisions on when to install priority lanes. This highlights the importance of considering mixed traffic environments in designing autonomous vehicle infrastructure and sets the stage for future advancements in urban traffic management.

A hierarchical intersection system control framework in mixed traffic conditions

Signalized intersections play a crucial role in urban road system and are a significant source of vehicle delays. Conventional intersection control methods struggle to cope with increasingly challenging traffic conditions. Fortunately, the development of artificial intelligence(AI), information communication technology(ICT) and connected and automated vehicle(CAV) technologies has brought new opportunities to enhance traffic performance at intersections. Based on these technologies, in the mixed traffic environment of CAVs and connected human-driven vehicles (CHVs), this paper proposes a hierarchical intersection system control framework, which adopts intersection level distributed control and corridor level centralized control. At the intersection level, considering the impact of CAV-dedicated lanes and using the CAV platoon and “1 + n” platoon consisting of one leading CAV and n following CHVs as control units, the mixed integer quadratic programming problem and optimal control problem are formulated to collaboratively optimize signal timing, lane settings and vehicle trajectories at isolated signalized intersections. At the corridor level, based on the optimization results of the intersection level and the critical path information, the critical path performance function and multi-stage optimal trajectory control formula are constructed to optimize the signal offsets and vehicle trajectories, which are passed to the intersection level, and ultimately minimize vehicle delays in the corridor. Simulation experiments and sensitivity analysis demonstrate the effectiveness and stability of the proposed framework in reducing vehicle delay and improving fuel consumption.

Safety-aware human-lead vehicle platooning by proactively reacting to uncertain human behaving

Human-Lead Cooperative Adaptive Cruise Control (HL-CACC) is regarded as a promising vehicle platooning technology in real-world implementation. By utilizing a Human-driven Vehicle (HV) as the platoon leader, HL-CACC reduces the cost and enhances the reliability of perception and decision-making. However, state-of-the-art HL-CACC technology still has a great limitation on driving safety due to the lack of considering the leading human driver’s uncertain behavior. In this study, a HL-CACC controller is designed based on Stochastic Model Predictive Control (SMPC). It is enabled to predict the driving intention of the leading Connected Human-Driven Vehicle (CHV). The proposed controller has the following features: (i) enhanced perceived safety in oscillating traffic; (ii) guaranteed safety against hard brakes; (iii) computational efficiency for real-time implementation. The proposed controller is evaluated on a PreScan&Simulink simulation platform. Real vehicle trajectory data is collected for the calibration of the simulation. Results reveal that the proposed controller: (i) improves perceived safety by 19.17 % in oscillating traffic; (ii) enhances actual safety by 7.76 % against hard brakes; (iii) is confirmed with string stability. The computation time is approximately 3.2 ms when running on a laptop equipped with an Intel i5-13500H CPU. This indicates the proposed controller is ready for real-time implementation.

Coordinated Ramp Metering Considering the Dynamics of Mixed-Autonomy Traffic

The introduction of connected autonomous vehicles may bring opportunities and challenges to traditional traffic control instruments, like ramp metering. This paper starts by constructing the fundamental diagram for mixed-autonomy traffic based on the car-following behaviors of both connected autonomous vehicles and human-driven vehicles. Then, analyses are performed on the main factors that influence the critical velocity, critical density, and road capacity under mixed-autonomy traffic. Two methods named COE-HERO and TRLCRM are developed to support the implementations of coordinated ramp metering for freeways with mixed-autonomy traffic. The COE-HERO method enhances the HERO method by incorporating a critical occupancy estimation module. Both COE-HERO and TRLCRM consider dynamic traffic flow parameters of mixed-autonomy traffic. The TRLCRM method is a reinforcement learning-based approach with a two-stage training framework, enabling it to adapt to varying mixed-autonomy demand scenarios. Extensive microscopic simulations show that the learning-based TRLCRM method can effectively alleviate bottleneck congestion and is robust to deal with various traffic scenarios. The COE-HERO method performs better than the HERO method, indicating the necessity of critical occupancy estimation in the implementations of coordinated ramp metering.

Modeling and Analysis of Mixed Traffic Networks with Human-Driven and Autonomous Vehicles

The emergence of connected and automated vehicles (CAV) indicates improved traffic mobility in future traffic transportation systems. This study addresses the research gap in macroscopic traffic modeling of mixed traffic networks where CAV and human-driven vehicles coexist. CAV behavior is explicitly included in the proposed traffic network model, and the vehicle number non-conservation problem is overcome by describing the approaching and departure vehicle number in discrete time. The proposed model is verified in typical CAV cooperation scenarios. The performance of CAV coordination is analyzed in road, intersection and network scenario. Total travel time of the vehicles in the network is proved to be reduced when coordination is applied. Simulation results validate the accuracy of the proposed model and the effectiveness of the proposed algorithm.

Developing a jam-absorption strategy for mixed traffic flow at signalized intersections using deep reinforcement learning

ABSTRACT Jam-absorption driving (JAD) can effectively prevent the generation and propagation of traffic oscillation. To alleviate the traffic congestion in the signalized intersection with mixed traffic flow, including human driving vehicles (HDVs) and connected and automated vehicles (CAVs), this study provides a jam-absorption driving strategy based on the traffic delay prediction of the mixed platoon under traffic congestion. An online traffic congestion prediction method with the objective of JAD is proposed and focuses on the leaving state of the trajectory to achieve fast capture of congestion features. Then, with real-time status and prediction information, we develop a Jam-absorption driving strategy based on a deep reinforcement learning (DRL) model to improve adaptability to the mixed traffic environment. The results show that this strategy can suppress more than 70% of traffic oscillations with excellent execution efficiency, improving traffic safety and efficiency.

Modeling and analysis of heterogeneous traffic flow with mixed regular and connected human-driven vehicles considering driver stochasticity

This paper presents a novel microscopic modeling framework to explore the impact of driver stochasticity and other related factors on heterogeneous traffic flows, consisting of both regular human-driven vehicles (RHDVs) and connected human-driven vehicles (CHDVs). We apply the stochastic optimal velocity car-following model for RHDVs, while an extended version of the stochastic continuous car-following model, taking into account optimal velocity guidance and driver compliance, is devised for CHDVs. The stability of CHDV car-following model is examined theoretically and numerically. Simulation experiments are conducted to analyze the stochastic heterogeneous traffic flow properties (i.e. stability and safety) with the proposed microscopic modeling framework under diverse parameter settings including CHDV penetration rate, driver compliance, and driver stochasticity. Results show that traffic flow stability and safety improve with the increase in CHDV penetration rate and driver compliance but deteriorate as the driver’s stochasticity strength increases. CHDVs significantly enhance traffic stability when the CHDV penetration rate or driver compliance exceeds a certain threshold, typically ranging between 0.5 and 0.7. Additionally, there is an interaction between driver stochasticity and CHDV penetration rate in their effect on traffic flow stability, where the contribution of stochasticity to traffic oscillations first increases and then decreases as the penetration rate rises. When the penetration rate is between 0 and 0.4, the traffic risk in emergency situations can be significantly reduced as the penetration rate increases. These findings may guide the management and control strategies of heterogeneous traffic flow.

String stable control design for automated vehicles under cyberattacks

Emerging automated vehicle (AV) technologies open a door for cyberattacks, where a select number of AVs are compromised to drive in an adversarial manner, degrading the performance of transportation systems. Hence, designing stable controls for AVs that can mitigate the impact of attacks on traffic flow is important and necessary as AVs gradually become a reality. In this study, we first present a general framework for describing mixed-autonomy traffic involving AVs and human-driven vehicles (HVs) based on car-following dynamics. Under this framework, a class of malicious attacks on AV control commands (vehicle acceleration) is mathematically characterized. To deal with the lack of knowledge of the attacks, we model attacks as an unstructured process, allowing for the inclusion of both deterministic and stochastic attack behaviors, bounded by a known bound (to remain stealthy) without being subject to any specific statistical distribution. Moreover, we analytically derive a set of sufficient conditions for string stable control design of individual AVs which can help mitigate the disturbances to traffic flow caused by unknown attacks on AVs. Furthermore, for any given market penetration rate of AVs, a set of conditions is also derived for selecting appropriate feedback gains of AVs to ensure string stability of heterogeneous traffic, reducing the undesired impact of attacks on traffic flow. These sufficient conditions provide important criteria for string stable control design of AVs without requiring much knowledge of the attacks. A series of numerical results is presented to show effectiveness of the proposed approach on mitigating attack disruption.

The effects of communication topology on mixed traffic flow: modelling and analysis

This paper aims to evaluate the combined impact of the communication topology (CT) and vehicular spatial distribution (SD) on mixed traffic of human-driven vehicles (HDVs) and connected and autonomous vehicles (CAVs). A generalized car-following (CF) model is presented to capture mixed traffic dynamics, integrating the effects of the CT, human reaction time, information delay and the optimal velocity changes with memory. Then, the linear stable condition of the model is derived using the perturbation method. Finally, extensive simulations are conducted under different CTs (i.e., multiple-predecessor leader following (MPLF) topology, predecessor following (PF) topology, and predecessor-leader following (PLF) topology) and six kinds of SDs. The results reveal that the MPLF topology demonstrates the best steady performance (i.e., stability) and dynamic performance (i.e., smoothness), and that the SD where CAVs are concentrated at the front of the traffic flow shows superior dynamic performance.

Wide human-like neural network incorporating driving styles for human-like driving intention analysis

Enhancing the synergy between autonomous and human-driven vehicles at the societal level requires understanding drivers’ behaviors and cognitive patterns, as well as conducting human-like driving intention analysis. To achieve this goal, this study designs a novel framework for analyzing human-like driving intention. Firstly, a spectral clustering method is employed to characterize driving styles. Secondly, a misclassification cost matrix is tailored to different driving needs. Finally, inspired by the complex neural networks found in the human brain, we develop a specialized lightweight neural network, termed the Width Human-like Neural Network (WNN), aimed at realizing personalized cognition and facilitating human-like decision-making in driving intention. Experimental studies and validation based on natural driving trajectory data from Kunming, China, demonstrate that the method accurately infers internal implicit driving intention from external explicit and observable driving behaviors, achieving a prediction accuracy of 99.8%. This framework strategically allocates limited computational resources to critical areas for autonomous vehicles and exemplifies best practices for improving neural network performance in driving intention analysis tasks.

Large language model as parking planning agent in the context of mixed period of autonomous vehicles and Human-Driven vehicles

Autonomous vehicles (AVs) are anticipated to revolutionize future transportation, necessitating updates to traffic infrastructure, particularly parking facilities, due to the unique characteristics of AVs compared to Human-Driven Vehicles (HDVs). During the transition period in which AVs and HDVs coexist, adaptable infrastructure is essential to accommodate both vehicle types. Traditional research, typically reliant on complex mathematical models and simulations, faces challenges in adapting to diverse urban settings, requiring substantial time and resources. To address these challenges, a government-level framework was developed, enabling urban planners to quickly and accurately evaluate and optimize existing parking facilities for future AV and HDV coexistence scenarios. The framework integrates a Large Language Model (LLM) to enhance flexibility and efficiency in parking planning throughout the transitional period. Structured guidance is incorporated to enhance decision-making precision and reduce LLM hallucination risks. The flexibility, robustness, and accuracy of the framework were validated through step-by-step and end-to-end testing using real-world datasets. Specifically, the framework achieved 91.1 % comprehensiveness and 70.2 % consistency in Indicator Selection Module testing, a 68.9 % success rate in the Single Indicator Calculation Module, and a 66.7 % success rate in end-to-end testing, demonstrating its practical value in supporting cities during AV integration. Finally, the success rates of different LLM agent modules were further explored, along with a comparison of multiple LLMs and an analysis of key issues related to LLM trustworthiness in urban planning applications. The research highlights the potential of LLMs in advancing urban planning processes and optimizing existing infrastructure, contributing to smarter and more adaptable urban environments.

A parameter privacy-preserving strategy for mixed-autonomy platoon control

It has been demonstrated that leading cruise control (LCC) can improve the operation of mixed-autonomy platoons by allowing connected and automated vehicles (CAVs) to make longitudinal control decisions based on the information provided by surrounding vehicles. However, LCC generally requires surrounding human-driven vehicles (HDVs) to share their real-time states, which can be used by adversaries to infer drivers’ car-following behavior, potentially leading to financial losses or safety concerns. This paper aims to address such privacy concerns and protect the behavioral characteristics of HDVs by devising a parameter privacy-preserving approach for mixed-autonomy platoon control. First, we integrate a parameter privacy filter into LCC to protect sensitive car-following parameters. The privacy filter allows each vehicle to generate seemingly realistic pseudo states by distorting the true parameters to pseudo parameters, which can protect drivers’ privacy in behavioral parameters without significantly influencing the control performance. Second, to enhance the reliability and practicality of the privacy filter within LCC, we first introduce an individual-level parameter privacy preservation constraint to the privacy filter, focusing on the privacy level of each individual parameter pair. Subsequently, we extend the current approach to accommodate continuous parameter spaces through a neural network estimator. Third, analysis of head-to-tail string stability reveals the potential impact of privacy filters in degrading mixed traffic flow performance. Simulation shows that this approach can effectively trade off privacy and control performance in LCC. We further demonstrate the benefit of such an approach in networked systems, i.e., by applying the privacy filter to a preceding vehicle, one can also achieve a certain level of privacy for the following vehicle.

Safety evaluation for mixed traffic flow of CAVs with different automation and connection levels

With the development of connected communication and automated driving technologies, connected automated vehicles (CAVs) are expected to gradually replace human-driven vehicles (HDVs), owing to their significant safety advantages. However, widespread adoption of CAVs will still require a considerable amount of time due to constraints such as technological development, legal regulations, and social acceptance. During the transitional period of CAVs’ proliferation, road traffic will comprise vehicles with different levels of connection and automation. These vehicles will differ in technology regarding communication, perception, decision-making, and other aspects, potentially posing severe challenges to mixed traffic flow safety. To address this issue, this paper proposes a mixed traffic flow model from both automation and connection perspectives to consider time delay and communication degradation. Four safety evaluation metrics are employed, and three-speed scenarios are set to investigate the influence of connection levels, automation levels, and their combined effects on mixed traffic flow safety. The result shows that (1) low automation enhances traffic safety in a non-connected environment, whereas medium and high automation decreases traffic safety. Conversely, in a connected environment, the role of automation levels exhibits the opposite pattern. (2) non-connected vehicles enhance traffic safety in a low-automated environment, while connected vehicles diminish it. However, connection level effects are reversed in medium and highly automated environments. (3) In connected and automated environments, traffic safety is increased by non-connected and low-automated vehicles, connected and medium automated vehicles, and connected and highly automated vehicles. However, other vehicle configurations decrease traffic safety. (4) Across various scenarios, the integration of connection and high automation proves beneficial for reducing traffic collision risks and improving traffic safety. In summary, this study provides theoretical support for managing and controlling mixed traffic flow with different automation and connection levels in the future.

Congested traffic patterns of mixed lattice hydrodynamic model combining the perceptual range differences with passing effect

With the aid of Vehicle-to-everything (V2X) technologies for network connectivity, connected autonomous vehicles (CAVs) can broaden the drivers' perceptual boundaries and receive a greater quantity of exogenous vehicle information, thereby governing the vehicle's acceleration information of the next moment. Nonetheless, constrained by contemporary communication networks and sophisticated vehicle control technology, the process of promoting CAVs is long-lasting, and throughout this stage of transition, both CAVs and human-driven vehicles (HDVs) will coexist on the road. Moreover, passing behaviour has received limited attention in the research on the traffic flow models despite being a fundamental microscopic driving behaviour. To bridge these gaps, we introduce the percentage ratios of CAVs into the lattice hydrodynamic model by integrating the perceptual range differences between two different types of vehicles with passing effects. Subsequently, the stability norm associated with the new model is ascertained by performing the linear stability analysis. When the stability condition is not achieved, we investigate the complex behaviour of the new model. The associated existing conditions and the modified Korteweg-de Vries (mKdV) equation are determined simultaneously. When the passing ratio is inadequate, no jam and kink jam make up the whole phase region; when the passing ratio surpasses the minimum, the initial unstable region can be segregated into two extra segments: the chaotic sub-region and the kink jam sub-region, and the density wave progressively transitions from being a kink-Bando traffic wave to a chaotic phase. The findings of the numerical experiment are consistent with with the theoretical derivation.

CosGCTFormer: An end-to-end driver state recognition framework

Driver state is critical for safety in mixed traffic environments with automated and human-driven vehicles. Recognizing and monitoring the driver’s state is essential to provide early warnings of abnormal driving performance. This paper proposes a driver state recognition framework called cosGCTFormer for end-to-end recognition of the driver’s emotional state and mental load. First, we employ an inverted residual convolution module to extract local features from physiological signal data. Second, the cosGCT module is introduced to extract global features. This module uses cos attention to address the high computational complexity of the Transformer in calculating attention weights. Additionally, it employs Gated Channel Transformation (GCT) to establish relationships between different channel dimensions, something the Transformer alone cannot do effectively. Finally, a simulated driving experiment was designed to collect data for training and testing the proposed method. Results demonstrate that our framework achieves higher accuracy and a more lightweight design compared to prevalent Convolution neural network-based (CNN-based) and Transformer-based methods. Additionally, the effectiveness of both modules was verified.

Adaptive Cruise Control under threat: A stochastic active safety analysis of sensing attacks in mixed traffic

Mixed traffic environments combining human-driven vehicles (HDVs) and those equipped with Adaptive Cruise Control (ACC) have already become prevalent. This study tackles the critical yet underexplored threat of sensing attacks, such as jamming and spoofing, on ACC systems. By applying stochastically calibrated ACC and HDV car-following models grounded in field data, we constructed an integrated and high-fidelity framework to simulate mixed traffic. This allows us to comprehensively analyze traffic safety risks enabled by surrogate safety measures, under various sensing attack scenarios and considering mechanisms for cyberattack detection and human intervention. Our findings highlight profound vulnerabilities in traffic safety from sensing attacks, with factors including stochastic driving behaviors, ACC penetration rates, and attack effectiveness. Through scenario-based sensitivity analyses, this research underscores the potential risks more realistically by stochastic simulation and also contributes to the design of detection systems to safeguard mixed traffic. Ultimately, this work provides valuable insights into evaluating the robustness of ACC systems against sensing attacks, supporting the ongoing and future development of effective countermeasures.

Analysis of mixed traffic flow characteristics based on cellular automata model under lane management measures

In the future, mixed traffic flow comprising Human-Driven Vehicles (HDVs), Connected and Autonomous Vehicles (CAVs), and CAV platoons will coexist for an extended period. Exploring the operational characteristics of mixed traffic flow under different lane management measures is essential for effective management and control. Initially, we analyze the motion characteristics of HDVs, CAVs, and CAV platoons and identify the car-following types within mixed traffic. Based on vehicle motion characteristics and the variance in maximum desired speeds among HDV drivers, we establish longitudinal motion rules for HDVs using a safety distance model. For CAVs and their platoons, we develop longitudinal motion rules by considering platoon merging and splitting behaviors, as well as speed and spacing error requirements for platoon driving. Subsequently, we formulate lateral lane-change rules based on the Symmetric Two-lane Cellular Automata (STCA) model, considering differences in reaction times and following distances between HDVs and CAVs. Finally, we conducted simulation experiments on a unidirectional three-lane highway using the multi-lane mixed traffic flow cellular automata model, analyzing the characteristics of mixed traffic under various lane management measures, such as mixed lane, CAV-dedicated lane, and CAV-priority lane. The results show that as the proportion of CAV (denoted as p) increases, the traffic flow capacity, optimal density, and jam density also increase. When p≤0.6, the k−q diagram exhibits a triangular shape, whereas for p>0.6, it assumes a trapezoidal shape. Implementing CAV-dedicated lane can reduce Cooperative Adaptive Cruise Control (CACC) degradation rates but only enhances traffic flow capacity when the CAV proportion reaches a certain threshold. Compared to the mixed lane scheme, when p≤0.3 and k≤0.3, the CAV-priority lane schemes not only meet traffic demands but also reduce CACC degradation rates. The vehicle speed in CAV-priority lane surpasses that in HDV lanes, facilitating improved traffic efficiency for CAVs. The distribution of maximum speeds among HDV drivers affects the fundamental diagram of mixed traffic flow and the performance of CAV-priority lane, with greater impacts observed as the standard deviation of the HDVs' maximum speed increases.

Multiclass Traffic Assignment Model Considering Heterogeneous Stochastic Headways of Autonomous Vehicles and Human-Driven Vehicles

Multiclass Traffic Assignment Model Considering Heterogeneous Stochastic Headways of Autonomous Vehicles and Human-Driven Vehicles

Navigating the blame game: Investigating automated vehicle fault in collisions under mixed traffic conditions

Vehicle being at fault in a crash has extensively been associated with its driver's behaviors and other human errors for human-driven vehicles (HDV). The introduction of automated vehicles (AVs) is expected to eliminate such human errors due to the ability of AVs to communicate with the external environment. However, various reports have documented AVs being at fault in collisions. This study applied text mining and mixed-effects logistic regression (MELR) on crash data involving AVs collected between 2017 and 2022 in California to explore the likelihood of an AV being at fault during a collision. It was found that among 497 crashes, a relatively small percentage (14.29 %) involved AVs being at fault. The text network results revealed patterns of keywords associated with the AVs being at fault. Such patterns include conventional mode of operation, area of impact, and resulting injuries. Furthermore, with about a 93 % prediction accuracy and an 83 % sensitivity score, the MELR results revealed that the likelihood of AVs being at fault increases when they are operated in conventional mode or when disengagement is involved. Moreover, turning, merging, or changing lane movements, unclear weather conditions, and operating on roadways with four or more lanes significantly increased the odds of an AV being at fault during a crash. Conversely, AVs were less likely to be at fault in commercial land use than residential land use, at intersection locations, and when the crash involved a truck. The practical implications of the findings are presented to improve AV operations.