Showing papers in "IEEE Intelligent Transportation Systems Magazine in 2020"
TL;DR: The architecture of various cooperative CAV systems is reviewed to answer how cooperative longitudinal motion control can work with the help of multiple system modules and what the critical design issues are.
Abstract: Connected and automated vehicles (CAVs) have the potential to address a number of safety, mobility, and sustainability issues of our current transportation systems. Cooperative longitudinal motion control is one of the key CAV technologies that allows vehicles to be driven in a cooperative manner to achieve system-wide benefits. In this paper, we provide a literature survey on the progress accomplished by researchers worldwide regarding cooperative longitudinal motion control systems of multiple CAVs. Specifically, the architecture of various cooperative CAV systems is reviewed to answer how cooperative longitudinal motion control can work with the help of multiple system modules. Next, different operational concepts of cooperative longitudinal motion control applications are reviewed to answer where they can be implemented in today's transportation systems . Different cooperative longitudinal motion control methodologies and their major characteristics are then described to answer what the critical design issues are . This paper concludes by describing an overall landscape of cooperative longitudinal motion control of CAVs, as well as pointing out opportunities and challenges in the future research and experimental implementations.
194 citations
TL;DR: A summary of the research efforts in shared autonomous vehicle systems that have been reported in the literature to date and discuss potential future research directions is presented in this article, where the implications of a system with shared autonomous vehicles have been investigated.
Abstract: Shared mobility can provide access to transportation on a custom basis without vehicle ownership. The advent of connected and automated vehicle technologies can further enhance the potential benefits of shared mobility systems. Although the implications of a system with shared autonomous vehicles have been investigated, the research reported in the literature has exhibited contradictory outcomes. In this paper, we present a summary of the research efforts in shared autonomous vehicle systems that have been reported in the literature to date and discuss potential future research directions.
72 citations
TL;DR: A framework for ground vehicle localization that uses cellular signals of opportunity (SOPs), a digital map, an inertial measurement unit (IMU), and a Global Navigation Satellite System (GNSS) receiver is developed to enable localization in an urban environment where GNSS signals could be unusable or unreliable.
Abstract: A framework for ground vehicle localization that uses cellular signals of opportunity (SOPs), a digital map, an inertial measurement unit (IMU), and a Global Navigation Satellite System (GNSS) receiver is developed. This framework aims to enable localization in an urban environment where GNSS signals could be unusable or unreliable. The proposed framework employs an extended Kalman filter (EKF) to fuse pseudorange observables extracted from cellular SOPs, IMU measurements, and GNSS-derived position estimates (when available). The EKF is coupled with a map-matching approach. The framework assumes the positions of the cellular towers to be known, and it estimates the vehicle's states (position, velocity, orientation, and IMU biases) along with the difference between the vehicle-mounted receiver clock error states (bias and drift) and each cellular SOP clock error state. The proposed framework is evaluated experimentally on a ground vehicle navigating in a deep urban area with a limited sky view. Results show a position root-mean-square error (RMSE) of 2.8 m across a 1,380-m trajectory when GNSS signals are available and an RMSE of 3.12 m across the same trajectory when GNSS signals are unavailable for 330 m. Moreover, compared to localization with a loosely coupled GNSS?IMU integrated system, a 22% reduction in the localization error is obtained whenever GNSS signals are available, and an 81% reduction in the localization error is obtained whenever GNSS signals are unavailable.
70 citations
TL;DR: A systematic procedure for lane detection based on the trajectories of vehicles collected on the road with the roadside LiDAR sensor that can release engineers from the manual lane identification task and can avoid any error caused by manual work.
Abstract: How to collect the real-time information of unconnected vehicles has been a challenge for connected vehicle technologies. The LiDAR sensors deployed along the roadside and at intersections provide a solution to fill the data gap during the transition from the traditional traffic to the full connected traffic. The roadside LiDAR sensors can record the movement of all road users with a relative long detection range. The lane detection serves as a fundamental but important step for LiDAR data processing. The location (which lane is occupied) of vehicles can be used for lane changing detection, vehicle departure warning and wrong-way alerts. But currently, there is not an effective method to identify the boundary of lanes using roadside LiDAR sensors. This paper presents a systematic procedure for lane detection based on the trajectories of vehicles collected on the road with the roadside LiDAR sensor. The whole procedure includes two major parts: background filtering and road boundary identification. This robust procedure can release engineers from the manual lane identification task and can avoid any error caused by manual work. Two case studies confirmed the effectiveness of the proposed method. Compared to previous lane detection methods, this procedure is not affected by the existence of pedestrians. This method can also detect the boundaries of lanes from the roads having curves with the limited time cost.
60 citations
TL;DR: The proposed algorithm can significantly improve the integrated complexity of the generated test scenarios while ensuring the coverage, which can help to find potential faults of the system more and faster, and further enhance the test efficiency.
Abstract: In this paper, we propose a new scenario generation algorithm called Combinatorial Testing Based on Complexity (CTBC) based on both combinatorial testing (CT) method and Test Matrix (TM) technique for intelligent driving systems. To guide the generation procedure in the algorithm and evaluate the validity of the generated scenarios, we further propose a concept of complexity of test scenario. CTBC considers both overall scenario complexity and cost of testing, and the reasonable balance between them can be found by using the Bayesian optimization algorithm on account of the black box property of CTBC. The effectiveness of this method is validated by applying it to the lane departure warning (LDW) system on a hardware-in-the-loop (HIL) test platform. The result shows that the bigger the complexity index is, the easier it is to reveal system defects. Furthermore, the proposed algorithm can significantly improve the integrated complexity of the generated test scenarios while ensuring the coverage, which can help to find potential faults of the system more and faster, and further enhance the test efficiency.
47 citations
TL;DR: This article investigates a traffic-flow forecasting problem based on long–short term memory (LSTM), an artificial recurrent neural network architecture used in deep learning, and presents a novel graph-attention LSTM structure, which leverages the strength of the graph-Attention mechanism for non-Euclidean structured data modeling and that of the L STM cell for time-series modeling.
46 citations
TL;DR: A novel magnetic tracking approach to improve the positioning accuracy of AGVs using a super strong magnetic nail instead of the low-remanence magnetic nail, which can be more easily tracked by a two-dimensional (2D) rather than 1 D sensor array.
Abstract: Automated guided vehicles (AGVs) have been widely adopted in the logistic delivering of modern manufacturing. As a key performance index for an AGV, the positioning accuracy of commercial AGVs based on the traditional magnetic tracking approach is bigger than ?5mm, which cannot meet the requirement of many industrial applications. Thus, we proposed a novel magnetic tracking approach to improve the positioning accuracy of AGVs. A super strong magnetic nail, instead of the low-remanence magnetic nail, can be more easily tracked by a two-dimensional (2D) rather than 1 D sensor array. The magnetic flux intensity around the magnetic nail can be expressed as a dipole model. Hence, the location and orientation of the nail can be computed via the sensor array data and a hybrid optimization algorithm, which is combined by the particle swarm optimization (PSO) algorithm and Levenberg-Marquardt (LM) algorithm performed on a microcontroller. We carried out experiments to verify the performance of the proposed positioning system in a series of initial driving speeds and target distances, where an N35 neodymium magnetic nail functioned as the designated AGV positioning point. Results show that the average positioning accuracy is improved to ?1.69mm, and the positioning accuracy can be further improved by a better motion control strategy. In addition, our proposed magnetic tracking approach can be easily fused with other navigation approaches such as laser and inertial sensing.
34 citations
TL;DR: A new method is presented that can optimize both operation cost and passenger quality of service in a global and farsighted view by leveraging the historical data by formulated as a Markov decision process considering idle vehicle rebalancing.
Abstract: This paper investigates a ride-sharing vehicle dispatching and routing problem in ride-sharing autonomous mobility-on-demand systems. We present a new method that can optimize both operation cost and passenger quality of service in a global and farsighted view by leveraging the historical data. It comprises two parts, one for vehicle routing decision making and the other for request-vehicle assignment. In particular, the vehicle routing decision making procedure is formulated as a Markov decision process considering idle vehicle rebalancing, with properly designed states, actions, and rewards. By sampling the future requests according to the historical probability distribution, the look-ahead decision making is realized via a deep reinforcement learning framework, which is composed of a Convolutional Neural Network and a double deep Q-learning module. Then a request-vehicle assignment scheme is presented based on the learning value attained from vehicle routing. Satisfactory performances (e.g, service rate, average waiting time and travel distance) of the method are demonstrated by experimental results under various fleet sizes and different vehicle capacities.
34 citations
TL;DR: Examining the traffic flow impacts of converting HOV lanes into CACC lanes regarding CACC MPRs on a complex freeway corridor with multiple interacting bottlenecks in California shows that converting to C ACC lanes at low M PRs can exacerbate congestion in general purpose lanes, whereas at mediate CACCMPRs the congestion is drastically alleviated due to a large share of traffic carried by CACC vehicles.
Abstract: Cooperative Adaptive Cruise Control (CACC) systems can increase roadway capacity, but the benefits are marginal at low market penetration rates (MPRs). Thus, a CACC dedicated lane is considered to group CACC vehicles for efficient traffic stream. Concepts of converting existing High Occupancy Vehicle (HOV) lanes into CACC lanes emerge, which leverages the infrastructural facilities and experience with HOV lanes. However, it is unclear to which extent changing HOV lanes to CACC lanes can influence freeway operations. This study examines the traffic flow impacts of converting HOV lanes into CACC lanes regarding CACC MPRs on a complex freeway corridor with multiple interacting bottlenecks in California. A simulation model capable of reproducing flow characteristics with HOV lane and CACC systems is employed for the assessment. Special attention is paid to macroscopic congestion patterns, CACC lane utilization, travel time reliability and CACC operation characteristics. The results show that converting to CACC lanes at low MPRs (130%) can exacerbate congestion in general purpose lanes, whereas at mediate CACC MPRs (40%-50%) the congestion is drastically alleviated due to a large share of traffic carried by CACC lanes.
30 citations
TL;DR: A comparison with existing unaided systems for urban GNSS processing indicates that the proposed system has significantly greater availability or accuracy, and a performance- sensitivity analysis reveals that navigation data wipeoff for fully modulated GNSS signals and a dense reference network are key to high-performance urban RTK positioning.
Abstract: This article presents the most thorough study to date of vehicular carrier-phase differential Global Navigation Satellite System (CDGNSS) positioning performance in a deep urban setting unaided by complementary sensors. Using data captured during approximately 2 h of driving in and around the dense urban center of Austin, Texas, a CDGNSS system is demonstrated to achieve 17-cm-accurate 3D urban positioning (95% probability) with a solution availability greater than 87%. The results are achieved without the aid of inertial, electro-optical, or odometry sensors. The development and evaluation of the unaided, GNSS-based precise positioning system is a key milestone toward the overall goal of combining precise GNSS, vision, radar, and inertial sensing for all-weather, high-integrity, high-absolute-accuracy positioning for automated and connected vehicles. The system described and evaluated herein is composed of a densely spaced reference network, a software-defined GNSS receiver, and a real-time kinematic (RTK) positioning engine. A performance- sensitivity analysis reveals that navigation data wipeoff for fully modulated GNSS signals and a dense reference network are key to high-performance urban RTK positioning. A comparison with existing unaided systems for urban GNSS processing indicates that the proposed system has significantly greater availability or accuracy.
29 citations
TL;DR: In this article, the authors proposed a mobility-on-demand (MOD) system consisting of shared autonomous vehicles (SAVs) to improve the efficiency of urban transportation through reduced vehicle ownership and parking demand.
Abstract: Mobility-on-demand (MOD) systems consisting of shared autonomous vehicles (SAVs) are expected to improve the efficiency of urban transportation through reduced vehicle ownership and parking demand. However, several issues related to their implementation remain open, such as unifying the vehicle and ridesharing (RS) assignment with rebalancing (RB) unoccupied vehicles.
TL;DR: This article focuses on two complementary spoofing-detection techniques available on commercial GPS receivers and thus require no additional hardware to operate, which are able to distinguish between interference and spoofing.
Abstract: Due to the ever-growing threat of GPS spoofing, it has become necessary for the aviation sector to develop an effective means of detection. This article focuses on two complementary spoofing-detection techniques that are available on commercial GPS receivers and thus require no additional hardware to operate. The primary methodology for detection is using a combination of radio power monitoring metrics, levering both automatic gain control and C/N0 measurements, along with multiple correlations for signal distortion to provide a best practices spoofing-detection algorithm, which is able to distinguish between interference and spoofing. The article first assesses nominal statistics for both metrics compiled from more than 250 h of nominal data collected from multiple wide area augmentation system stations. These data are compared to previous collections to validate the thresholds and false alarms rates and establish a complete testing methodology. These tests and thresholds are then assessed with the Texas spoofing test battery series of GPS spoofing data sets to confirm detection capabilities. Finally, these tests and thresholds are applied to assess the GPS signal of six extended flights over the United States to assess the performance on an aircraft.
TL;DR: The role of PNT information for enabling inland waterway navigation-assistance functions, such as bridge-collision warning, mooring aiding, and automatic guidance is discussed, and the real-time kinematic technique is developed to provide integrity information alongside reliable and centimeter-level-accurate PNT data.
Abstract: Inland navigation and shipping are important pillars of the European Transport System. To support the skipper during safety-critical operations, precise position, navigation, and timing (PNT) data are required. This work discusses the role of PNT information for enabling inland waterway navigation-assistance functions, such as bridge-collision warning, mooring aiding, and automatic guidance. The real-time kinematic technique is developed to provide integrity information alongside reliable and centimeter-level-accurate PNT data. In addition, the transmission of Global Navigation Satellite Systems (GNSS) correction data is investigated. Since GSM does not currently meet the availability and stability communication requirements along inland waterways, use of the automatic identification system (AIS) for data transmission is explored. The proposed navigation solution and the communication developments were analyzed in real time in challenging GNSS signal-degraded scenarios on the authorized inland waterway testbed. Despite the further developments required for the AIS communication infrastructure, it is shown that our system architecture can nearly meet the integrity and accuracy requirements for driver-assistance functions on inland waterways.
TL;DR: It is clear that a consistent solution for ethical challenges does not currently exist and recommendations are provided, to potentially encourage the involving players to collaborate to tackle the remaining challenges.
Abstract: Ethical issues remain a significant concern for future large-scale deployment of autonomous vehicles. Although machine ethics have been developed for a few decades, ethical decision making in autonomous driving imposes more complex and emerging challenges to which some dedicated efforts from academia, policymakers, and automakers have been devoted in the past few years. This paper reviews the efforts and progress associated with the various aspects of ethical challenges in autonomous vehicles. Based on the critical review, it is clear that a consistent solution for ethical challenges does not currently exist and recommendations are provided, to potentially encourage the involving players to collaborate to tackle the remaining challenges.
TL;DR: Experimental results indicate that the proposed GNSS?lidar integration can obtain improved positioning accuracy in a highly urbanized area in Hong Kong.
Abstract: Positioning is a key function for autonomous vehicles that requires globally referenced localization information. Lidarbased mapping, which refers to simultaneous localization and mapping (SLAM), provides continuous positioning in diverse scenarios. However, SLAM error can accumulate through time. Besides, only relative positioning is provided by SLAM. The Global Navigation Satellite System (GNSS) receiver is one of the significant sensors for providing globally referenced localization, and it is usually integrated with lidar in autonomous driving. However, the performance of the GNSS is severely challenged due to the reflection and blockage caused by buildings in superurbanized cities, including Hong Kong, China; Tokyo; and New York, resulting in the notorious non-line-of-sight (NLOS) receptions. Moreover, the uncertainty of the GNSS positioning is ambiguous, leading to the incorrect tuning of its weight during GNSS?lidar integration. This article innovatively employs lidar to identify the NLOS measurement of the GNSS receiver using point-cloud-based object detection. Measurements from satellites suffering from NLOS reception will be excluded based on the proposed fault detection and exclusion (FDE) algorithm. Then, GNSS-weight least-square positioning is conducted based on the surviving measurements from FDE. The noise covariance of the GNSS positioning is calculated by considering the potential location errors caused by the NLOS and the remaining LOS measurements. The improved GNSS result and its corresponding noise covariance are integrated with lidar through a graph-based SLAM-integration framework. Experimental results indicate that the proposed GNSS?lidar integration can obtain improved positioning accuracy in a highly urbanized area in Hong Kong.
TL;DR: This article proposes two integrity monitoring schemes with two classes of approaches, based on either the snapshot weighted least-squares residual or sequential weighted extended Kalman filter innovation, which can effectively improve positioning accuracy and guarantee system integrity.
Abstract: Global Navigation Satellite System (GNSS) integrity is defined as a measure of trust that can be placed in the correctness of the information supplied by the total system. Initially developed for safetycritical applications in the aeronautic domain, this concept has attracted more and more attention from terrestrial GNSS-based applications in recent years. The main problem of integrity monitoring for urban transport applications is related to GNSS signal degradation due to local effects such as multipath and non-line-of-sight reception. This article proposes two integrity monitoring schemes with two classes of approaches, based on either the snapshot weighted least-squares residual or sequential weighted extended Kalman filter innovation. The integrity monitoring schemes are mainly realized by two modules: accuracy enhancement, in which measurement errors are better characterized, and integrity monitoring, in which one of the fault detection and exclusion techniques, i.e., the Danish reweighting method, is applied. The results with real GPS data collected in urban canyons show that the proposed system can effectively improve positioning accuracy and guarantee system integrity.
TL;DR: The articles in this special section focus on recent advancements on the use of the global navigation satellite system (GNSS)-based positioning for intelligent transport systems.
Abstract: The articles in this special section focus on recent advancements on the use of the global navigation satellite system (GNSS)-based positioning for intelligent transport systems. The civil applications of geopositioning are undergoing exponential development. The latest market analysis for global navigation satellite systems (GNSSs) shows two major fields of application which share the majority of the market: intelligent transport systems (ITS), mainly in the road ITS domain, and location-based services, accessible on smartphones and tablets. The modernization of GPS and Russia’s GLONASS system and the development of Galileo and Bei- Dou are proceeding at a fast pace, introducing improved potential capabilities and higher performance levels for satellite-based positioning, and leading to new architectures for positioning and new strategies for positioning by means of other sensors. GNSSs are considered the superior system to provide accurate and global position, velocity, and time.
TL;DR: In this paper, the role and potential of these paradigms in the context of V2X communication is discussed, and how softwarization, virtualization, and machine learning can be adapted to tackle the challenges of such networks.
Abstract: The concept of the fifth generation (5G) mobile network system has emerged in recent years as telecommunication operators and service providers look to upgrade their infrastructure and delivery modes to meet the growing demand. Concepts such as softwarization, virtualization, and machine learning will be key components as innovative and flexible enablers of such networks. In particular, paradigms such as software-defined networks, software-defined perimeter, cloud & edge computing, and network function virtualization will play a major role in addressing several 5G networks' challenges, especially in terms of flexibility, programmability, scalability, and security. In this work, the role and potential of these paradigms in the context of V2X communication is discussed. To do so, the paper starts off by providing an overview and background of V2X communications. Then, the paper discusses in more details the various challenges facing V2X communications and some of the previous literature work done to tackle them. Furthermore, the paper describes how softwarization, virtualization, and machine learning can be adapted to tackle the challenges of such networks.
TL;DR: The reformulated punctuality problem can be further transformed into the standard form of integer linear programming (ILP), thus, can be efficiently solved by using the existing ILP solvers.
Abstract: This paper focuses on a specific stochastic shortest path (SSP) problem, namely the punctuality problem. It aims to determine a path that maximizes the probability of arriving at the destination before a specified deadline. The popular solution to this problem always formulates it as a cardinality minimization problem by considering its data-driven nature, which is approximately solved by the 1 , -norm relaxation. To address this problem accurately, we consider the special character in the cardinality-based punctuality problem and reformulate it by introducing additional variables and constraints, which guarantees an accurate solution. The reformulated punctuality problem can be further transformed into the standard form of integer linear programming (ILP), thus, can be efficiently solved by using the existing ILP solvers. To evaluate the performance of the proposed solution, we provide both theoretical proof of the accuracy, and experimental analysis against the baselines. Particularly, the experimental results show that in the following two scenarios, 1) artificial road network with simulated travel time, 2) real road network with real travel time, our accurate solution works better than others regarding the accuracy and computational efficiency. Furthermore, three ILP solvers, i.e., CBC, GLPK and CPLEX, are tested and compared for the proposed accurate solution. The result shows that CPLEX has obvious advantage over others.
TL;DR: Vision and lidar are complementary sensors that are incorporated into many applications of intelligent transportation systems and the validation of algorithm robustness has recently been recognized as a major challenge for the massive deployment of these new technologies.
Abstract: Vision and lidar are complementary sensors that are incorporated into many applications of intelligent transportation systems. These sensors have been used to great effect in research related to perception, navigation, and deep learning applications. Despite this success, the validation of algorithm robustness has recently been recognized as a major challenge for the massive deployment of these new technologies.
TL;DR: A rolling-horizon scheme that dynamically optimizes taxi dispatching considering the actual traffic conditions, and a clustering-based technique that can significantly improve the computation time without harming the solution quality is introduced.
Abstract: Taxis are an important transportation mode in many cities due to their convenience and accessibility. In the taxi-dispatching problem, sometimes it is more beneficial for the supplier if taxis cruise in the network after serving the first request to pick up the next passenger, while sometimes it is better that they wait in stations for new trip requests. In this article, we propose a rolling-horizon scheme that dynamically optimizes taxi dispatching considering the actual traffic conditions. To optimize passenger satisfaction, we define a limitation for passenger waiting time. To be able to apply the method to large-scale networks, we introduce a clustering-based technique that can significantly improve the computation time without harming the solution quality. Finally, we test our method on a real test case considering taxi requests with personal car trips to reproduce actual network loading and unloading congestion during peak hours.
TL;DR: A big data perspective for investigating the DDB based on models, data features, and experiences is presented and a new perspective was presented for researchers, applicants or business managers to do more effective and suitable studies in the field of big data DDB.
Abstract: There are many articles published in driving/driver behavior (DDB) but few of them have focused on DDB with big data in the literature. The reasons for this might be the lack of media coverage, data, expertise or big data perspectives. This paper presents a big data perspective for investigating the DDB based on models, data features, and experiences. For this purpose, DDB studies were reviewed and grouped into six perspectives. 6V's of big data (volume, velocity, variety, veracity, vulnerability and value) were also revised and discussed how these V's were compatible with DDB data. Finally, the use of big data in DDB analysis was discussed and some suggestions were presented. The lack of big data perception on DDB research was also overviewed and a new perspective was presented for researchers, applicants or business managers to do more effective and suitable studies in the field of big data DDB.
TL;DR: A multiagent simulation model is presented to observe and assess the effects of real-time information provision on the passengers in transit networks and shows that real- time personalized information may have an increasingly positive impact on overall travel times following the increasing ratio of connected passengers.
Abstract: Modeling and simulation play an important role in transportation networks analysis. With the widespread use of personalized real-time information sources, the status of the simulation depends heavily on individual travelers reactions to the received information. As a consequence, it is relevant for the simulation model to be individual-centered, and agent-based simulation is the most promising paradigm in this context. Information is now personalized, and the simulations have to take into account the interaction of individually guided passengers. In this paper, we present a multiagent simulation model to observe and assess the effects of real-time information provision on the passengers in transit networks. These effects are measured by simulating several scenarios according to the ratio of connected passengers to a real-time information system. We represent the passengers and the vehicles as agents in the system. We analyze the simulated scenarios following their effect on the passengers travel times. The information provided to the connected passengers is based on a space-time representation of the transportation networks. Results show that real-time personalized information may have an increasingly positive impact on overall travel times following the increasing ratio of connected passengers. However, there is a ratio threshold after which the effect of real-time information becomes less positive.
TL;DR: This paper proposes for the first time the use of an Internet of Things solution for the accurate recovery of incapacitated commercial and retail delivery drones that have crashed either due to technical issues or outside malign intervention.
Abstract: This paper proposes for the first time the use of an Internet of Things solution for the accurate recovery of incapacitated commercial and retail delivery drones. Since the use of drones will increase in popularity for a variety of uses, the problem of locating and recovering delivery drones is necessary. Using the emerging LoRaWAN and Sigfox networks (as examples of Low Power Wide Area Network (LPWAN) technology), we investigate the opportunity to locate and recover delivery drones that have crashed either due to technical issues or outside malign intervention. Location and recovery of a test drone using LoRa technology was trialed over three distinct terrains (urban, suburban, and rural) at five different sites in each terrain. These 15 trials evaluate ease of recovery on arrival at the crash site, accuracy of location coordinates given, and a subsequent analysis of the channel statistics for the LoRa network. The experiment was repeated using the Sigfox network as a comparison. The drone was able to be recovered at 14 of the 15 tests sites for both LoRa and Sigfox; in every case the drone's location was successfully transmitted to the secure server via the LPWAN networks. The paper investigates the reported location accuracy from the drone and also uses RMSE as an accuracy metric. The paper furthermore divulges lessons learned and presents a drone recovery algorithm.
TL;DR: A lane keeping simulator that is built with image projections of recorded data in conjunction with vehicle dynamics estimation and an end-to-end learning method using convolutional neural network takes front-view camera data as input and produces the proper steering wheel angle to keep the vehicle in lane.
Abstract: Autonomous lane keeping is an important safety feature for intelligent vehicles. This paper presents a lane keeping simulator that is built with image projections of recorded data in conjunction with vehicle dynamics estimation. An end-to-end learning method using convolutional neural network (CNN) takes front-view camera data as input and produces the proper steering wheel angle to keep the vehicle in lane. A novel method of data augmentation is proposed using vehicle dynamic model and vehicle trajectory tracking, which can create additional training data as if a vehicle drives off-lane in any displacement and orientation. Experimental results demonstrate that the CNN model trained with the simulator can achieve higher accuracy for autonomous lane keeping and much lower failure rate. The simulator can serve as a platform for both training and evaluation of vision based autonomous driving algorithms. The experimental dataset is made available at http://computing.wpi.edu/dataset.html .
TL;DR: It is found that the fusion of Global Navigation Satellite System (GNSS) and inertial navigation systems (INSs) does not provide the accuracy required for automated driving, and the use of optical sensors is challenged by snow covering the road markings.
Abstract: This article provides an overview of the use of inertial and visual sensors and discusses their prospects in the Arctic navigation of autonomous vehicles. We also examine the fusion algorithms used thus far for integrating vehicle localization measurements as well as the map-matching (MM) algorithms relating position coordinates with road infrastructure. Our review reveals that conventional fusion and MM methods are not enough for navigation in challenging environments, like urban areas and Arctic environments. We also offer new results from testing inertial and optical sensors in vehicle positioning in snowy conditions. We find that the fusion of Global Navigation Satellite System (GNSS) and inertial navigation systems (INSs) does not provide the accuracy required for automated driving, and the use of optical sensors is challenged by snow covering the road markings. Although extensive further research is needed to solve these problems, the fusion of GNSS, INSs, and optical sensors seems to be the best option due to their complementary nature.
TL;DR: Experimental results show that the proposed multistep LSTM-NN model outperforms all the benchmark models, especially in a commercial parking lot with heavy traffic flow.
TL;DR: This study uses mobile phone data to understand mobility patterns in a country, with limited mobility data, in order to give advice about decisions on how to design the national and regional road network.
Abstract: This study uses mobile phone data to understand mobility patterns in a country, with limited mobility data, in order to give advice about decisions on how to design the national and regional road network. Our method consists of three parts: (1) filtering mobile phone traces to derive mobility patterns, (2) building an adapted formulation of the gravity-based trip distribution model, which considers telecommunication intensity (i.e., aggregate number of calls and text messages) and travel time as input to forecast the influence of road improvements on country-wide mobility, and (3) optimizing the road network investment based on the adapted trip distribution model by using a local search algorithm. The method was applied to the case study country of Senegal. The mobile phone data was transformed to support informed decisions on road network development in that country given different objectives, namely accessibility and equity. We believe that the methodology is valuable and reproducible to other countries where traditional mobility data is scarce but mobile phone data is available to transport planners.
TL;DR: The main contribution of this paper is the design of a realtime risk-sensitive nonlinear model predictive controller to plan the online cost-effective cruising velocity in a stochastic traffic environment and closed-loop Entropic Value-at-Risk as a coherent risk measure is introduced.
Abstract: Battery Electric Vehicles have a high potential in modern transportation, however, they are facing limited cruising range. The driving style, the road geometries including slopes, curves, the static and dynamic traffic conditions such as speed limits and preceding vehicles have their share of energy consumption in the host electric vehicle. Optimal energy management based on a semi-autonomous ecological advanced driver assistance system can improve the longitudinal velocity regulation in a safe and energy-efficient driving strategy. The main contribution of this paper is the design of a realtime risk-sensitive nonlinear model predictive controller to plan the online cost-effective cruising velocity in a stochastic traffic environment. The basic idea is to measure the relevant states of the electric vehicle at runtime, and account for the road slopes, the upcoming curves, and the speed limit zones, as well as uncertainty in the preceding vehicle behaviour to determine the energy-efficient velocity profile. Closed-loop Entropic Value-at-Risk as a coherent risk measure is introduced to quantify the risk involved in the system constraints violation. The obtained simulation and field experimental results demonstrate the effectiveness of the proposed method for a semi-autonomous electric vehicle in terms of safe and energy-efficient states regulation and constraints satisfaction.
TL;DR: Simulation results show that the proposed method can save up to 18% fuel compared to linear quadratic (LQ) controller, and all vehicles achieve bounded range error and velocity fluctuation.
Abstract: The platooning of connected and automated vehicles has the potential to significantly improve the fuel efficiency of road transportation. Shortening the carfollowing distance to reduce aerodynamic drag is often used to improve fuel economy in today's platoons, but suffers high risk of rear-end collision. This paper presents an alternative solution to reduce platoon fuel consumption, i.e., periodic longitudinal control, which remains effective under the condition of middle or long following distance. The periodic controller can be applied to heterogeneous platoons, in which the optimal switching rule is designed for each vehicle to maintain a safe following distance while guaranteeing a collective bounded stability. Sectionalized switching maps based on state trajectory analysis are designed to choose the appropriate driving mode. The bounded stability for heterogeneous platoons is proved through the set-invariance theory, which ensures that the inter-vehicle distances are confined in a desirable range. Simulation results show that the proposed method can save up to 18% fuel compared to linear quadratic (LQ) controller, and all vehicles achieve bounded range error and velocity fluctuation.