Virtual Immersive Reality for Stated Preference Travel Behavior Experiments: A Case Study of Autonomous Vehicles on Urban Roads

Transportation Research Part C: Emerging Technologies

2020

Self Cite

Blockchain has the potential to render the transaction of information more secure and transparent. Nowadays, transportation data are shared across multiple entities using heterogeneous mediums, from paper collected data to smartphone. Most of this data are stored in central servers that are susceptible to hacks. In some cases shady actors who may have access to such sources, share the mobility data with unwanted third parties. A multi-layered Blockchain framework for Smart Mobility Data-market (BSMD) is presented for addressing the associated privacy, security, management, and scalability challenges. Each participant shares their encrypted data to the blockchain network and can transact information with other participants as long as both parties agree to the transaction rules issued by the owner of the data. Data ownership, transparency, auditability and access control are the core principles of the proposed blockchain for smart mobility data-market. In a case study of real-time mobility data sharing, we demonstrate the performance of BSMD on a 370 nodes blockchain running on heterogeneous and geographically-separated devices communicating on a physical network. We also demonstrate how BSMD ensures the cybersecurity and privacy of individual by safeguarding against spoofing and message interception attacks and providing information access management control. Service 1 that "...(Waze will) share personal information with companies or organizations connected or affiliated with Waze...". So all the information collected by Waze could be shared with other organizations associated with Waze. However, the Terms of Service do not make it clear if the user can track with whom their information has been shared.Another point related to privacy is the access control of the information that is provided by the users. It was recently discovered that Google keeps collecting user location data even if they explicitly deactivate the tracking system in their mobiles [26]. Another example of disclosing information without the user's consent is the infamous Facebook-Cambridge Analytica scandal [9].The General Data Protection Regulation 2 (GDPR) is a step forward for privacy. Although the GDPR is valid only in the European Union, it is still expected to push multinational companies to be more transparent on how they manage people's private information. In the authors opinion, no matter how many rules and fines governments apply, given the current centralized ways of collecting people's data, some entities will always find a loophole in legislation or hackers will be able to tamper with these centralized data systems. In addition to laws, individuals have to have total control and ownership of their information. We need to be the guardians and responsible of our own privacy, unfortunately nowadays we have to trust third parties for that.Distributed ledger technologies like blockchain have the potential to give the people full control of their information, protect individual's personal mobility information and guard their privacy. The ...

Section: Identification Layermentioning

confidence: 99%

A multi-layered blockchain framework for smart mobility data-markets

López

Transportation Research Part C: Emerging Technologies

2020

Self Cite

“…vehicles, cyclists, etc.). [2] This platform has recently been used to capture pedestrians' movements in a dense urban environment under varying traffic conditions as well as their general interaction and behaviour towards AVs. [13] In the scenarios, respondents stand at the side of a two-lane road with vehicles driving by from one or both directions.…”

Section: Deep Reinforcement Learningmentioning

confidence: 99%

Multi-Objective Autonomous Braking System using Naturalistic Dataset

Vásquez

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

2019

Self Cite

A deep reinforcement learning based multi-objective autonomous braking system is presented. The design of the system is formulated in a continuous action space and seeks to maximize both pedestrian safety and perception as well as passenger comfort. The vehicle agent is trained against a large naturalistic dataset containing pedestrian road-crossing trials in which respondents walked across a road under various traffic conditions within an interactive virtual reality environment. The policy for brake control is learned through computer simulation using two reinforcement learning methods i.e. Proximal Policy Optimization and Deep Deterministic Policy Gradient and the efficiency of each are compared. Results show that the system is able to reduce the negative influence on passenger comfort by half while maintaining safe braking operation.Another important aspect of transferable autonomous braking systems is having a realistic formulation of the problem during training. This includes accurately representing both the objectives and the mechanism used to reach those objectives. Intuitively, the braking control mechanism of a vehicle resides within a continuous (real-valued) action space. By avoiding the discretization of the action space, we can consider more complex driving behaviours which capture reality more accurately, such as the consideration of passenger comfort.While the pedestrian safety is a top priority, passenger comfort is an important and often overlooked factor in the acceptance and successful implementation of AVs [1]. Naturalness and perceived safety of the AV's driving manoeuvres are two major influences on a rider's comfort. Thus, designing an objective function which accounts for this is crucial for training an acceptable AV braking system.In our work, we demonstrate the use of highly realistic pedestrian road-crossing data collected by the Virtual Immersive Reality Environment (VIRE) [2] platform to train a multi-objective continuous autonomous braking system in simulation.

“…ReliefF Top n Covariates Figure 1: DeepWait framework In this model, R is the vector of independent variables, χ is the vector of parameters to be estimated, and ξ(t) is the underlying hazard [17]. Equation (2) gives the risk at time t for pedestrian i, where the underlying hazard expresses hazard or risk for a pedestrian at all time points regardless of the independent variables (R=0). To estimate model parameters, partial derivatives of the log-likelihood function is taken.…”

Section: All Covariates Rankeld Covariatesmentioning

confidence: 99%

“…Virtual Immersive Reality Environment (VIRE) [2] was used in this experiment to simulate an unsignalized crosswalk with a mixed-traffic. Before each experiment, participants were asked to fill a form on their sociodemographic information, walking habits, health conditions, and previous VR experiments.…”

Section: Data Collectionmentioning

confidence: 99%

“…AVs yielding to crossing pedestrians in a pedestrian-friendly urban area raises the necessity for new approaches in studying pedestrian-vehicle interactions [1], which emphasizes on the advantages of mid-block unsignalized crossings.In an attempt to study pedestrians' crossing behaviour, in this paper we investigate the pedestrian waiting time before crossing mid-block unsignalized crosswalks in fully automated, mixed-traffic, and fully human-driven conditions. As it is impossible to collect real data for studies on future technologies and their impact, and using questionnaires and stated preferences surveys may result in unrealistic answers [2], an immersive Virtual Reality (VR) based 3D environment was designed for the purpose of this study. Participants were asked to cross a simulated crosswalk in different scenarios, which were empowered by theoretical designs of experiments.…”

mentioning

confidence: 99%

See 1 more Smart Citation

DeepWait: Pedestrian Wait Time Estimation in Mixed Traffic Conditions Using Deep Survival Analysis

Kalatian

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

2019

Self Cite

Pedestrian's road crossing behaviour is one of the important aspects of urban dynamics that will be affected by the introduction of autonomous vehicles. In this study we introduce DeepWait, a novel framework for estimating pedestrian's waiting time at unsignalized mid-block crosswalks in mixed traffic conditions. We exploit the strengths of deep learning in capturing the nonlinearities in the data and develop a cox proportional hazard model with a deep neural network as the log-risk function. An embedded feature selection algorithm for reducing data dimensionality and enhancing the interpretability of the network is also developed. We test our framework on a dataset collected from 160 participants using an immersive virtual reality environment. Validation results showed that with a C-index of 0.64 our proposed framework outperformed the standard cox proportional hazard-based model with a C-index of 0.58.Extensive presence of autonomous vehicles (AVs) in urban areas is going to be one of the major changes in near future. AVs can fundamentally alter dynamics of urban areas. One of the aspects of urban mobility that can be affected by the widespread adaption of AVs is pedestrian's crossing behaviour. Studying this behaviour is important for the safety of most vulnerable users, optimal roadway layout changes, geometric design updates, and traffic flow optimization. AVs yielding to crossing pedestrians in a pedestrian-friendly urban area raises the necessity for new approaches in studying pedestrian-vehicle interactions [1], which emphasizes on the advantages of mid-block unsignalized crossings.In an attempt to study pedestrians' crossing behaviour, in this paper we investigate the pedestrian waiting time before crossing mid-block unsignalized crosswalks in fully automated, mixed-traffic, and fully human-driven conditions. As it is impossible to collect real data for studies on future technologies and their impact, and using questionnaires and stated preferences surveys may result in unrealistic answers [2], an immersive Virtual Reality (VR) based 3D environment was designed for the purpose of this study. Participants were asked to cross a simulated crosswalk in different scenarios, which were empowered by theoretical designs of experiments. In the meantime, participant's reactions and behaviours, , i.e. their coordinates, head orientations, stress level, etc. were being recorded, adding up to their sociodemographic data collected by questionnaires before the experiments.Using the collected data, a cox proportional hazard model is developed to estimate pedestrian's waiting time based on different factors. To better capture nonlinearities involved in the data, we then proposed DeepWait, a deep neural network based cox proportional hazard model, backed up by a feature selection algorithm to select most important factors affecting pedestrian waiting time.The rest of this paper is organized as follows: In the next section, an overview of the literature on pedestrian crossing behaviour, and particularly waiting time is pro...