Browse Topic: Automated vehicles

Items (375)
ABSTRACT Over time, the National Institute of Standards and Technology (NIST) has refined the 4Dimension / Real-time Control System (4D/RCS) architecture for use in Unmanned Ground Vehicles (UGVs). This architecture, when applied to a fully autonomous vehicle designed for missions in urban environments, can greatly assist in the process of saving time and lives by creating a more intelligent vehicle that acts in a safer and more efficient manner. Southwest Research Institute (SwRI®) has undertaken the Southwest Safe Transport Initiative (SSTI) aimed at investigating the development and commercialization of vehicle autonomy as well as vehicle-based telemetry systems to improve active safety systems and autonomy. This paper will discuss the implementation of the 4D/RCS architecture to the SSTI autonomous vehicle, a 2006 Ford Explorer.
McWilliams, GeorgeBrown, Michael
New forms of highly automated Advanced Air Mobility (AAM) aircraft, such as electric vertical take-off and landing (eVTOL) vehicles, could transform transportation, cargo delivery, and a variety of public services. The National Aeronautics and Space Administration (NASA) conducted a series of flight demonstrations in collaboration with the Defense Advanced Research Projects Agency (DARPA) and Sikorsky Aircraft (a Lockheed Martin company) to progressively evaluate autonomous technologies. The autoland flight test research is a first in series for investigating the world’s first procedural descending-decelerating automated landing with vertical guidance Instrument Flight Procedures (IFP). The Sikorsky Optionally Piloted Vehicle (OPV) experimental UH-60 Black Hawk was used to evaluate a flight path’s four-dimensional trajectory (4DT) management into primitive commands and then follow those commands to a Point-in-Space (PinS) landing to the ground. All flight procedures were manually flown to the ground at 12 degrees with a 20-knot tail wind to ensure flight safety before automation was engaged. New and novel high precision approach procedures could pave the way for all future VTOL operations.
Zahn, DavidPatterson, GayleWilliams, EthanEggum, SarahFettrow, Tyler
Abstract In this article, we present a spatiotemporal trajectory planning algorithm for emergency obstacle avoidance. Utilizing obstacle and driving environment data from the sensing module, we construct a 3D spatiotemporal grid map. This informs our improved hybrid A* algorithm, which identifies collision-safe, dynamically feasible trajectories. The traditional hybrid A* algorithm is enhanced in three significant ways to make the search practical and feasible: (1) optimizing search efficiency with motion primitives based on child node acceleration, (2) integrating collision risk into the heuristic function to reduce ineffective node exploration, and (3) introducing a One-Shot search based on the Optimal Boundary Value Problem (OBVP) to improve goal state searches. Finally, the algorithm is tested in two scenarios: (1) a vehicle cut-in from an adjacent lane and (2) a pedestrian crossing. Simulation results indicate that our proposed emergency obstacle avoidance trajectory planning method can efficiently devise trajectories that not only circumvent obstacles safely and adhere to vehicle dynamics constraints, but also meet the real-time demands of emergency obstacle avoidance trajectory planning.
Chen, GuoyingYao, JunGao, ZhenhaiGao, ZhengZhao, XuanmingXu, NanHua, Min
In this work, we present a lightweight pipeline for robust behavioral cloning of a human driver using end-to-end imitation learning. The proposed pipeline was employed to train and deploy three distinct driving behavior models onto a simulated vehicle. The training phase comprised of data collection, balancing, augmentation, preprocessing, and training a neural network, following which the trained model was deployed onto the ego vehicle to predict steering commands based on the feed from an onboard camera. A novel coupled control law was formulated to generate longitudinal control commands on the go based on the predicted steering angle and other parameters such as the actual speed of the ego vehicle and the prescribed constraints for speed and steering. We analyzed the computational efficiency of the pipeline and evaluated the robustness of the trained models through exhaustive experimentation during the deployment phase. We also compared our approach against state-of-the-art implementation in order to comment on its validity.
Samak, Tanmay VilasSamak, Chinmay VilasKandhasamy, Sivanathan
Validating an Approach to Assess Sensor Perception Reliabilities Without Ground Truth2021-01-00804/6/2021
A reliable environment perception is a requirement for safe automated driving. For evaluating and demonstrating the reliability of the vehicle’s environment perception, field tests offer testing conditions that come closest to the vehicle’s driving environment. However, establishing a reference ground truth in field tests is time-consuming. This motivates the development of a procedure for learning the vehicle’s perception reliability from fleet data without the need for a ground truth, which would allow learning the perception reliability from fleet data. In Berk et al. (2019), a method based on Bayesian inference to determine the perception reliability of individual sensors without the need for a ground truth was proposed. The model utilizes the redundancy of sensors to learn the sensor’s perception reliability. The method was tested with simulated data. In this contribution, we further explore and validate the method by utilizing real data, including ground truth data based on high-resolution LIDAR and human labeling. An area with overlapping field of view from five sensors is selected for the analysis. A basic association method is used to compare the object data obtained from the different sensors. Finally, we compare the sensor perception reliabilities learned from the Bayesian inference model with the sensor perception reliabilities determined from the labeled ground truth. In this paper, it is shown that the model introduced in Berk et al. (2019) can approximate the reference data based on the provided ground truth. The estimated parameters of the model do not perfectly correspond to the sensor reliabilities but are of the same order of magnitude as when derived from the ground truth.
Kryda, MarcoBerk, MarioBuschardt, BorisStraub, Daniel
Dynamically Adjustable LiDAR with SPAD Array and Scanner2021-01-00914/6/2021
An important function of an Automated Driving (AD) system is to detect objects including vehicles and pedestrians on the road. Typical devices for detecting those objects include cameras, millimeter-wave RADAR, and light detection and ranging (LiDAR). LiDAR uses the flight time of a short-wavelength electromagnetic wave. Because of that LiDAR is expected to find even small objects such as tire fragments on a road in high resolution. The detection performance required for LiDAR depends on the operational design domain (ODD). For example, while a vehicle is travelling at high speeds, LiDAR needs to detect apparently small objects at long distances, and while it is travelling at low speeds, LiDAR has to detect objects over a wide angular range. Conventional LiDAR is developed to satisfy all requirements, providing performance including detection distance, resolution, and angle of view tends to expose issues such as cost and size when it is mounted onboard. To solve these problems, we have built LiDAR with a new structure consisting of an originally developed light receiving unit and scanning unit, which are the main components. The light receiving units uses an array of high-sensitivity single-photon avalanche diodes (SPADs). Its vertical resolution can be selected by changing the number of SPADs per pixel. The scanning unit has introduced a reciprocal motion system, which enables dynamically choosing the range and speed of scanning, with the range of scanning 100 ° or wider. With these mechanisms, it is possible to select a high-resolution and narrow-angle mode when detecting small objects at long distances, and low-resolution and wide-angle mode for detecting many objects at short distances. Therefore, the LiDAR can adjust its performance dynamically according to driving scenes. We have confirmed that our LiDAR is effective for detecting objects under various conditions.
Nakajima, MasatoHata, TakehiroUeno, AkifumiOzaki, NoriyukiMizuno, FumiakiKashiwada, ShinjiYanai, Kenichi
Driving Automation System Test Scenario Development Process Creation and Software-in-the-Loop Implementation2021-01-00624/6/2021
Automated driving systems (ADS) are one of the key modern technologies that are changing the way we perceive mobility and transportation. In addition to providing significant access to mobility, they can also be useful in decreasing the number of road accidents. For these benefits to be realized, candidate ADS need to be proven as safe, robust, and reliable; both by design and in the performance of navigating their operational design domain (ODD). This paper proposes a multi-pronged approach to evaluate the safety performance of a hypothetical candidate system. Safety performance is assessed through using a set of test cases/scenarios that provide substantial coverage of those potentially encountered in an ODD. This systematic process is used to create a library of scenarios, specific to a defined domain. Beginning with a system-specific ODD definition, a set of core competencies are identified. These core competencies are then considered both in isolation and in conjunction with other potential confounding factors (e.g. other traffic or atmospheric conditions); with “edge cases” being represented as compounded or unique sets of confounding factors. Using this approach, a candidate scenario set is presented, along with a discussion of nuances and necessary considerations in scenario selection. These approaches are combined in a simulated environment to demonstrate their use. Finally, a strategy is proposed to automate the overall scenario testing process to make the execution less cumbersome. This process of test scenario creation strictly follows the ISO 26262 concept phase to verify the safety goals and functional safety requirements.
Patil, MayurLybarger, AlexanderMidlam-Mohler, ShawnStoddart, Evan
System Architecture Design Suitable for Automated Driving Vehicle: Hardware Configuration and Software Architecture Design2021-01-00734/6/2021
Our L2-automated driving system enabling a driver to take his/her hands off from the steering wheel is self-operating on a highway, allowing the vehicle to automatically change lanes and overtake slow-speed leading vehicles. It includes an OTA function, which can extend the ODD after the market launch. To realize these features in reasonably safer and more reliable ways, system architecture must be designed well under hardware and software implementation constraints. One such major constraint is the system must be designed to make the most out of the existing sensor configuration on the vehicle, where five peripheral radars and a front camera for ADAS as well as panoramic-view and rear-view cameras for monitoring are available. In addition, four LiDARs and a telephoto camera are newly adopted for ADS. Another constraint is the system must consist of reliable redundant components for fail-safe operation. When one component is dysfunctional due to a malfunction or temporal system limitation, others must keep the functionalities properly so that vehicle control is maintained without loss of sight of the surrounding objects for at least four seconds so that the driver can reasonably safely resume manual driving. To attain a high-level of backup controllability and visibility, we employ the redundant sensor configuration along with a redundant battery, steering and break actuators, and communication channels. We have also designed the ECU where a high-performance SoCs and highly-reliable MCUs are mounted to achieve ASIL-D. A further constraint is the limited computational resources for embedded systems. Our ADS constitutes many functions: localization, object recognition, map-based lane generation, path planning, vehicle controlling, backup vehicle controlling, and HMI. It is designed so that each function can take advantage of the CPU core exclusively as much as possible so as to be less disrupted by other working functions.
Kunisa, AkiomiNemoto, YusukeKato, HiroshiHasegawa, TomonoriKato, MasanoriMashima, Tomohisa
Stability Criteria for Accurate Path Tracking in Automated Guided Vehicle Systems2021-01-00934/6/2021
Satisfactory performance of intelligent vehicles systems in tracking a predefined trajectory requires an efficient control scheme to generate steering control signals from posture errors (i.e., errors in position and orientation). For such systems, it is necessary at each instant to control steering action so that any deviation from the path is corrected in a stable manner, in a reasonable time and without any oscillation about the desired path. This paper deals with stability of motion and motion control of intelligent vehicle systems. In this regard, the general control structure and specification of an optimum range of predefined control parameters for accurate path tracking of these systems are determined. A two degree of freedom (DOF) nonlinear dynamic model is developed to represent their plane motion. Path tracking of the vehicle is attained by controlling the position and orientation errors about the trajectory, which is accomplished by modifying the steering input signal on the basis of error feedbacks to the controller. Employment of various stability criteria and other constraints such as applying the physical limits of the vehicle to the controlled system narrows down the range of control parameters, within which the controlled system would remain stable. Experimental results demonstrating the performance of the system are reported.
Mehrabi, Mostafa
Letter from the Guest Editors
Hamid, Umar Zakir AbdulSandblom, FredrikHabibovic, AzraLi, Bin
csp1071 test Construction of The Traffic Law and Regulation Framework for Automated DrivingSAE-PP-002902/18/2021
Road automated driving as a new generation of information technology and the integration of the transport industry's development has become a new round of global scientific and technological innovation and industrial transformation. This technology will promote the continuous upgrading of the field of road traffic. At present, the government, enterprises, and investors all take this as the goal and direction to accelerate automated driving in China. A reasonable traffic law and regulation system is required to promote the healthy development of automatic driving and fully release scientific and technological innovation subjects' vitality. Until now, China has formulated an official rule concerning the road testing of automated driving. According to this rule, no passenger and freight transportation can be officially applied using automatic vehicles. Therefore, this paper tries to construct the traffic law and regulation framework to promote automated driving development. This paper firstly summarizes the profound reform of automated driving on the road traffic industry in terms of vehicles, infrastructure, practitioners, and transportation services. Based on China's current administrative rules and traffic regulations, the legal and institutional obstacles in automatic driving are analyzed. This paper proposes a traffic law and regulation framework to promote the development of automated driving. The proposed framework can help the transport authority administrate automated driving in legal identity, demonstration application, transportation operation, practitioner management, and scene management. Finally, the policy suggestions to help develop automated driving are put forward. Through enhanced supervision, mutual recognition of qualification, regulation mode innovation, ecosystem construction, and multi-party cooperation, the automated driving market and partners' vitality can be significantly stimulated
SintzADMIN, JeneaneAnthony, Lindsay
7.0.108 - Challenges Faced for Parameterization & Validation of a Small Gasoline Engine Plant Model for Application of EMS DevelopmentSAE-PP-002842/4/2021
Control algorithm development for typical Engine Management System is a challenging task. To develop a reliable control algorithm, proper closed loop testing environment is required. In such development activity, it is of prime importance to validate the algorithm on a standalone target engine. This can be achieved in engine test cell where the actual engine will be controlled by prototype ECU. But this process has drawbacks like higher testing cost, time consuming, non-reusability of test bed etc. Simulation based engine plant model development for closed loop ECU testing is an effective technique for such application. Various generic engine models are available for such application.to suit a particular target engine these model need to be parameterized with precise engine data. The vehicle parameters used for parameterization are typically obtained from actual test and engine design data. This paper elaborates the process and challenges faced while parameterization and validation of engine plant model in simulation environment and steps followed while parameterizing a two cylinder gasoline engine to suit EMS development. Simulations were carried out in Model In Loop (MIL) and Hardware In Loop (HIL). Validation results were compared with actual vehicle data from dynamometer trials and are presented in corresponding sections.
Lname, Fname
7.0.107 - Design and Development of Capacitance Type Level Sensor for Automotive Vehicle ApplicationSAE-PP-002832/4/2021
Fuel level sensor is a device to indicate the level of the fuel in fuel tank fitted in an automobile. This will have features to communicate the fuel level to the dashboard of the vehicle and is of significant attention to the driver during vehicle usage. The advanced instrumentation provides a lot of information on the dashboard display such as information about fuel level, computing mileage, miles to go or miles to empty, fuel economy, average mileage, etc. Presently, the float arm type with Thick Film Resistor(TFR) and Reed switch type fuel level sensors are being used. To have accurate information for computing, the present sensors are not supporting due to its limitations like nonlinearity, fluctuating output due to slosh, output variations in steps and not continuous. The measurement accuracy of the fuel level sensor needs to be focused to rely on the information available on the dashboard instrument. Hence, it is vital to have a sensor with better reliability, accuracy and adaptability. Capacitance technology based level sensor is identified for development as one of the solutions to meet the demands of the above requirements and this paper portrays the complete perspective and design methodology of capacitance based fuel level sensor. The basics of capacitance measurement, design concepts, design validations, proto-type results are elaborated in this paper.
Lname, Fname
Construction of The Traffic Law and Regulation Framework for Automated DrivingSAE-PP-002462/3/2021
Road automated driving as a new generation of information technology and the integration of the transport industry's development has become a new round of global scientific and technological innovation and industrial transformation. This technology will promote the continuous upgrading of the field of road traffic. At present, the government, enterprises, and investors all take this as the goal and direction to accelerate automated driving in China. A reasonable traffic law and regulation system is required to promote the healthy development of automatic driving and fully release scientific and technological innovation subjects' vitality. Until now, China has formulated an official rule concerning the road testing of automated driving. According to this rule, no passenger and freight transportation can be officially applied using automatic vehicles. Therefore, this paper tries to construct the traffic law and regulation framework to promote automated driving development. This paper firstly summarizes the profound reform of automated driving on the road traffic industry in terms of vehicles, infrastructure, practitioners, and transportation services. Based on China's current administrative rules and traffic regulations, the legal and institutional obstacles in automatic driving are analyzed. This paper proposes a traffic law and regulation framework to promote the development of automated driving. The proposed framework can help the transport authority administrate automated driving in legal identity, demonstration application, transportation operation, practitioner management, and scene management. Finally, the policy suggestions to help develop automated driving are put forward. Through enhanced supervision, mutual recognition of qualification, regulation mode innovation, ecosystem construction, and multi-party cooperation, the automated driving market and partners' vitality can be significantly stimulated. Keywords: Traffic Law and Regulation, Automated Driving, Administrative Rules and regulations, Road Testing, Demonstration Application, Commercial Operation
MobrxivNonAdmin, Lindsay
1.1.216 - Tailored Design and Layout for Loss Minimization or Cost-Effective Commonality of Parts - A Contradictory ConflictSAE-PP-002422/3/2021
In order to minimize the development and production costs in the automotive industry, despite steadily increasing variety of models and applications offered by the OEMs, the pressure on standardization of components and production processes is increasing continuously. As a direct consequence, modular engine families are already established with high degrees of common parts and kits as well as standardized interfaces for all vehicle platforms by most manufacturers these days. At the same time, the world adopted and announced massive legal demands concerning the reduction of CO2 emissions for the entire vehicle fleet. In addition to the optimization of the combustion process, the exhaust gas aftertreatment and thermal management, the use of improved and more resilient materials for higher reduction of mechanical friction leads to a significant amount of the realized lowering in fuel consumption respective CO2 emissions. Significant future potential for friction reduction and loss minimization is expected to result from, for example, one for the particular application optimized, on-demand component dimensioning and tailored calibrations. This dedicated fine-tuning approach is contrary to the widely spread application of a clear common part strategy. In the course of this paper the question will be discussed whether the additional cost of a component diversification can be justified within an engine family in contradiction to a best cost approach by using the scaling effects of parts communization.
Mutagaana, Festo
2.0.104 - Tackling Three Critical Issues of Transportation: Environment, Safety and Congestion Via Semi-autonomous PlatooningSAE-PP-002142/1/2021
In recent years, platooning emerged as a realistic configuration for semi-autonomous driving. In the SARTRE project, simulation and physical tests were performed to validate the platooning system not only in testing facilities but also in conventional highways. Five vehicles were adapted with autonomous driving systems to have platooning functionalities, enabling to perform platoon tests and assess the feasibility, safety and benefits. Although the tested system was in a prototype, it demonstrated sturdiness and good functionality, allowing performing conventional road tests. First of all the fuel consumption decreased up to 16% in some configurations and different gaps between the vehicles were tested in order to establish the most suitable for platooning in terms of safety and economy. Additionally, the platooning technology enables a new level of safety in highways. Around 85% of the accident causation is the human factor. With platooning, the human factor is reduced almost entirely, making it a more efficient and safer system. With platooning, the traffic flow is more homogeneous with several vehicles travelling at the same speed and therefore accident situations are less probable while helping to maintain a steady traffic flow with no stop and go situations. For the same reason, the traffic flow is more manageable enabling advanced traffic management. Thus, the traffic congestion can be reduced and, again, reducing another source of emissions. All the potential advantages of a platooning technology which has been developed and tested are studied in this paper through test results and simulations.
Mutagaana, Festo
Automated vehicles require some level of subsystem redundancy, whether to allow a transition time for driver re-engagement (L3) or continued operation in a faulted state (L4+). Highly automated vehicle developers need to have safe miles accumulated by vehicles to assess system maturity and experience new environments. This article presents a conceptual framework suggesting that hardware newly available to commercial vehicle application can be used to form a steering system that will remain operational upon a failure. The key points of a provisional safety case are presented, giving hope that a complete safety case is possible. This article will provide autonomous vehicle developers a view of a near term possibility for a highly automated commercial vehicle steering solution.
Pandy, AnandaPathuri, NagamaniSalunke, PranavSubba, Srujana SreeWilliams, Dan
AVSC Best Practice for Passenger-Initiated Emergency Trip InterruptionAVSC000032020066/30/2020
As passengers take rides in fleet-managed automated driving system-dedicated vehicles (ADS-DVs), they may feel the need to interrupt the trip due to a perceived emergency. There is currently no industry consensus on the proper balance between ADS passenger agency and the potential for introducing unexpected outcomes in dynamic traffic environments. In order to build public trust in automated vehicles, passengers should be given an option to exercise some type of control (agency) to intervene during situations they perceive as emergencies. Passenger-initiated emergency trip interruption features — however they manifest in a given vehicle - can help establish this confidence in ADS technologies. AVSC Best Practice for Passenger-Initiated Emergency Trip Interruption recommends processes surrounding aspects of passenger-initiated features in SAE level 4 and 5 fleet-managed ADS-DVs. It recommends criteria and processes for passenger initiation of these features from inside the vehicle; communication with passengers and fleet operations; enhanced diagnoses of the situation, interaction outside the vehicle with other road users, and general post-stop actions. Also, precautions against some types of foreseeable misuse are addressed. These recommendations apply to commercially available, deployed ADS-DV’s providing trips to people. The AVSC recommends that every fleet-managed SAE level 4 and 5 ADS-DV be equipped with a (PES) or (PEC) or both.
Automated Vehicle Safety Consortium
AVSC Best Practice for Describing an Operational Design Domain: Conceptual Framework and LexiconAVSC000022020044/15/2020
An ADS-operated vehicle’s operational design domain (ODD) is defined by the manufacturer based on numerous factors. Research is underway at other organizations to define and organize ODD elements into taxonomies and other relational constructs. In order to enhance collaboration and communication between manufacturers and developers and transportation authorities, common terms and consistent frameworks are needed. The conceptual framework presented by Automated Vehicle Safety Consortium establishes a lexicon that can be used consistently by ADS developers and manufacturers responsible for defining their ADS ODD. A common framework and lexicon will reduce confusion, align expectations, and therefore build public trust, acceptance, and confidence. The guidance in this document is intended for: The technical community (e.g. manufacturers and developers) Public agencies (e.g. regulatory authorities) Infrastructure owner-operators The public This document, Best Practice for Describing an Operational Design Domain: Conceptual Framework and Lexicon is a critical first step. It offers a conceptual framework for manufacturers and developers to use when communicating with public agencies and the general public about their ADS’s ODD. It also details a list of potential variables with definitions that manufacturers and developers might use to describe certain aspects of the ODDs of their ADS-operated vehicles. It was developed with fleet-managed, SAE Level 4 vehicles in mind — i.e. vehicles requiring no human intervention to operate within their ODD. These vehicles are NOT privately owned.
Automated Vehicle Safety Consortium
Modes of Automated Driving System Scenario Testing: Experience Report and Recommendations2020-01-12044/14/2020
With the widespread development of automated driving systems (ADS), it is imperative that standardized testing methodologies be developed to assure safety and functionality. Scenario testing evaluates the behavior of an ADS-equipped subject vehicle (SV) in predefined driving scenarios. This paper compares four modes of performing such tests: closed-course testing with real actors, closed-course testing with surrogate actors, simulation testing, and closed-course testing with mixed reality. In a collaboration between the Waterloo Intelligent Systems Engineering (WISE) Lab and AAA, six automated driving scenario tests were executed on a closed course, in simulation, and in mixed reality. These tests involved the University of Waterloo’s automated vehicle, dubbed the “UW Moose”, as the SV, as well as pedestrians, other vehicles, and road debris. Drawing on both data and the experience gained from executing these test scenarios, the paper reports on the advantages and disadvantages of the four scenario testing modes, and compares them using eight criteria. It also identifies several possible implementations of mixed-reality scenario testing, including different strategies for data mixing. The paper closes with twelve recommendations for choosing among the four modes.
Antkiewicz, MichałKahn, MaximilianAla, MichaelCzarnecki, KrzysztofWells, PaulAcharya, AtulBeiker, Sven
Driving Safety Performance Assessment Metrics for ADS-Equipped Vehicles2020-01-12064/14/2020
The driving safety performance of automated driving system (ADS)-equipped vehicles (AVs) must be quantified using metrics in order to be able to assess the driving safety performance and compare it to that of human-driven vehicles. In this research, driving safety performance metrics and methods for the measurement and analysis of said metrics are defined and/or developed. A comprehensive literature review of metrics that have been proposed for measuring the driving safety performance of both human-driven vehicles and AVs was conducted. A list of proposed metrics, including novel contributions to the literature, that collectively, quantitatively describe the driving safety performance of an AV was then compiled, including proximal surrogate indicators, driving behaviors, and rules-of-the-road violations. These metrics, which include metrics from on- and off-board data sources, allow the driving safety performance of an AV to be measured in a variety of situations, including crashes, potential conflicts, and near misses. These measurements enable the evaluation of temporal flows and the quantification of key aspects of driving safety performance. The identification and exploration of metrics focusing explicitly on AVs as well as proposing a comprehensive set of metrics is a unique contribution to the literature. The objective is to develop a concise set of metrics that allow driving safety performance assessments to be effectively made and that align with the needs of both the ADS development and transportation engineering communities and accommodate differences in cultural/regional norms. Concurrent project work includes equipping an intersection with a sensor suite of cameras, LIDAR, and RADAR to collect data requiring off-board sources and employing test AVs to collect data requiring on-board sources. Additional concurrent work includes development of artificial intelligence and computer vision-based algorithms to automatically calculate the metrics using the collected data. Future work includes using the collected data and algorithms to finalize the list of metrics and then develop a methodology that uses the metrics to provide an overall driving safety performance assessment score for an AV.
Wishart, JeffreyComo, StevenElli, MariaRusso, BrendanWeast, JackAltekar, NirajJames, Emmanuel
Analysis of LiDAR and Camera Data in Real-World Weather Conditions for Autonomous Vehicle Operations2020-01-00934/14/2020
Autonomous vehicle technology has the potential to improve the safety, efficiency, and cost of our current transportation system by removing human error. With sensors available today, it is possible for the development of these vehicles, however, there are still issues with autonomous vehicle operations in adverse weather conditions (e.g. snow-covered roads, heavy rain, fog, etc.) due to the degradation of sensor data quality and insufficiently robust software algorithms. Since autonomous vehicles rely entirely on sensor data to perceive their surrounding environment, this becomes a significant issue in the performance of the autonomous system. The purpose of this study is to collect sensor data under various weather conditions to understand the effects of weather on sensor data. The sensors used in this study were one camera and one LiDAR. These sensors were connected to an NVIDIA Drive Px2 which operated in a 2019 Kia Niro. Two custom scenarios (static and dynamic objects) were chosen to collect sensor data operating in four real-world weather conditions: fair, cloudy, rainy, and light snow. An algorithm developed herein was used to provide a method of quantifying the data for comparison against the other weather conditions. The results from these performance algorithms show that sensor data quality degrades by an average of 13.88% for static objects and 16.16% for dynamic objects while operating in these conditions, with operations in rain proving to have the most significant effect on sensor data degradation. From this study, it is hypothesized that advancements in data processing algorithms can improve the usability of this degraded data. In future work, we seek to explore fault-tolerant sensor fusion algorithms that can overcome the effects of adverse weather.
Goberville, NickEl-Yabroudi, MohammadOmwanas, MarkRojas, JohanMeyer, RickAsher, ZacharyAbdel-Qader, Ikhlas
Data-Driven Framework for Fuel Efficiency Improvement in Extended Range Electric Vehicle Used in Package Delivery Applications2020-01-05894/14/2020
Extended range electric vehicles (EREVs) are a potential solution for fossil fuel usage mitigation and on-road emissions reduction. The use of EREVs can be shown to yield significant fuel economy improvements when proper energy management strategies (EMSs) are employed. However, many in-use EREVs achieve only moderate fuel reduction compared to conventional vehicles due to the fact that their EMS is far from optimal. This paper focuses on in-use rule-based EMSs to improve the fuel efficiency of EREV last-mile delivery vehicles equipped with two-way Vehicle-to-Could (V2C) connectivity. The method uses previous vehicle data collected on actual delivery routes and machine learning methods to improve the fuel economy of future routes. The paper first introduces the main challenges of the project, such as inherent uncertainty in human driver behavior and in the roadway environment. Then, the framework of our practical physics-model guided data-driven approach is introduced. For vehicles with small amounts of prior data, a Bayesian method is used to adjust a control parameter in the EMS offline for each vehicle with introduced prior information derived from large numbers of trips from other vehicles in the fleet. For vehicles with many delivery trips, a reinforcement learning algorithm is used to optimize the parameter in real-time without requiring future information of the trip. Although our data-driven framework cannot achieve a globally optimal solution with respect to the fuel efficiency, it provides a systematic and immediate solution for in-use EREVs used for package delivery with a very low computation cost and no change of vehicle hardware. Also, this framework is ready to be extended for further fuel economy improvements if more information is available from advanced transportation infrastructures like Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) connectivity.
Wang, PengyueNorthrop, William
A Connected Controls and Optimization System for Vehicle Dynamics and Powertrain Operation on a Light-Duty Plug-In Multi-Mode Hybrid Electric Vehicle2020-01-05914/14/2020
This paper presents an overview of the connected controls and optimization system for vehicle dynamics and powertrain operation on a light-duty plug-in multi-mode hybrid electric vehicle developed as part of the DOE ARPA-E NEXTCAR program by Michigan Technological University in partnership with General Motors Co. The objective is to enable a 20% reduction in overall energy consumption and a 6% increase in electric vehicle range of a plug-in hybrid electric vehicle through the utilization of connected and automated vehicle technologies. Technologies developed to achieve this goal were developed in two categories, the vehicle control level and the powertrain control level. Tools at the vehicle control level include Eco Routing, Speed Harmonization, Eco Approach and Departure and in-situ vehicle parameter characterization. Tools at the powertrain level include PHEV mode blending, predictive drive-unit state control, and non-linear model predictive control powertrain power split management. These tools were developed with the capability of being implemented in a real-time vehicle control system. As a result, many of the developed technologies have been demonstrated in real-time using a fleet of four instrumented Chevrolet Volts which are equipped with on-board sensors, rapid prototyping embedded controllers, and V2X communication devices. This paper provides an overview of each tool developed, its implementation, energy reduction in isolation, and the net energy reduction of various tool combinations. A breakdown of the energy savings and range extension possible for the connected vehicle control and optimization tool set is provided which shows energy reduction benefits approaching 20% and range extension upwards of 8%, dependent on the driving and traffic scenarios and initial vehicle state of charge.
Oncken, JosephOrlando, JoshuaBhat, Pradeep K.Narodzonek, BrandonMorgan, ChristopherRobinette, DarrellChen, BoNaber, Jeffrey
Joint Calibration of Dual LiDARs and Camera Using a Circular Chessboard2020-01-00984/14/2020
Environmental perception is a crucial subsystem in autonomous vehicles. In order to build safe and efficient traffic transportation, several researches have been proposed to build accurate, robust and real-time perception systems. Camera and LiDAR are widely equipped on autonomous self-driving cars and developed with many algorithms in recent years. The fusion system of camera and LiDAR provides state-of the-art methods for environmental perception due to the defects of single vehicular sensor. Extrinsic parameter calibration is able to align the coordinate systems of sensors and has been drawing enormous attention. However, differ from spatial alignment of two sensors’ data, joint calibration of multi-sensors (more than two sensors) should balance the degree of alignment between each two sensors. In this paper, we assemble a test platform which is made up of dual LiDARs and one monocular camera and use the same sensing hardware architecture as intelligent sweeper designed by our laboratory. Meanwhile, we propose the related joint calibration method using a circular chessboard. The center of circular chessboard is respectively detected in camera image to get pixel coordinates and in point cloud of LiDAR to get 3D coordinates. The calibration problem is then converted into a 3D-2D PnP matching problem and the center of the chessboard is set as corresponding points to construct the geometric constraints to get initial calibration values. Further, a proper global loss function is elaborately designed for Levenberg-Marquardt nonlinear optimization to obtain the final calibration parameters, and then the extrinsic parameters between any two sensors are estimated simultaneously. Experimental results show that the proposed method is suitable for the joint calibration of fusion system composed of LiDARs and camera, and the calibration results have high accuracy and stability.
Deng, ZhenwenXiong, LuYin, DongShan, Fengwu
Perceptions of Two Unique Lane Centering Systems: An FOT Interview Analysis2020-01-01084/14/2020
The goal of this interview analysis was to explore and document the perceptions of two unique lane centering systems (S90’s Pilot Assist and CT6’s Super Cruise). Both systems offer a similar type of functionality (adaptive cruise control and lane centering), but have significantly different design philosophies and HMI (Human-Machine Interface) implementations. Twenty-four drivers drove one of the two vehicle models for a month as part of a field operational test (FOT) study. Upon vehicle return, drivers took part in a 60-minute semi-structured interview covering their perceptions of the vehicle’s various advanced driver-assistance systems (ADAS). Transcripts of the interviews were coded by two researchers, who tagged each statement with relevant system and perception code labels. For analysis, the perception codes were grouped into larger thematic bins of safety, comfort, driver attention, and system performance. Perceptions of adaptive cruise control (ACC) were similar across vehicles. Almost all participants mentioned benefits of comfort and safety associated with ACC use. Participants cited different benefits between the two vehicle’s implementations of lane centering. A majority of participants (75%) described comfort benefits associated with Super Cruise, while less than half (41%) cited comfort benefits associated with Pilot Assist. Only a few participants (25%) mentioned safety benefits associated with Super Cruise. Half (50%) of the participants mentioned safety benefits associated with Pilot Assist. Almost all participants cited fears of potential misuse of the system in which drivers might pay less attention to the driving task. Results suggest that drivers’ comprehension and expectation of these systems’ behavior are strongly influenced by their design philosophies, specifically in terms of the difference in hands-on versus hands-off-wheel implementation. The perceived role of the driver – as either a fallback driver or as an assisted driver - may be influenced by the design implementation.
Landry, StevenSeppelt, BobbieRusso, LucaMehler, BruceAngell, LindaGershon, PninaReimer, Bryan
Intention-aware Lane Changing Assistance Strategy Basing on Traffic Situation Assessment2020-01-01274/14/2020
Traffic accidents avoidance is one of the main advantages for automated vehicles. As one of the main causes of vehicle collision accidents, lane changing of the ego vehicle in case that the obstacle vehicles appear in the blind spot with uncertain motion intentions is one of the main goals for the automated vehicle. An intention-aware lane changing collision assistance strategy basing on traffic situation assessment in the complex traffic scenarios is proposed in this paper. Typical Regions of Interest (ROI) within the detection range of the blind spots are selected basing on the road topology structures and state space consisting of the ego vehicle and the obstacle vehicles. Then the motion intentions of the obstacle vehicles in ROI are identified basing on Gaussian Mixture Models (GMM) and the corresponding motion trajectories are predicted basing on the state equation. Traffic situation is assessed according to the index of the motion intentions and the coupling tendency between the ego vehicle and the obstacle vehicles and the risk level is graded basing on the map with collision time. Lane keeping assist is carried out according to the assessment result of the traffic situation. Testing scenarios with the straight road and T-junction are designed and a co-simulation environment consisting of CarMaker and Mathwork Simulink is established to verify the proposed strategy in complex traffic scenes. Simulation results present an adaptive ROI and a high identification accuracy for motion intentions of the obstacle vehicles. What’s more, it shows that the traffic situation can be accurately evaluated and the ego vehicle can be effectively controlled with the appearance of the high-risk vehicles.
Wu, JianLiu, SihanHe, RuiSun, Bohua
The Design of Safe-Reliable-Optimal Performance for Automated Driving Systems on Multiple Lanes with Merging Features2020-01-01224/14/2020
Safety function for automated driving systems including advanced driver assistance systems and autonomous vehicle systems is very important. Inside safety function, predictive judge sub-function should be designed with the consideration of more and more penetration of automated driving vehicles. This paper presents the design on multiple lanes with merging features based on the author's previous Patent JP2019-147944 using predictive time-head-way and time-to-collision maps. In the author's previous work (Model Predictive Control for Hybrid Electric Vehicle Platooning Using Slope Information-Published on IEEE Transactions on Intelligent Transportation Systems), a model predictive control framework was designed. Due to the difficulty to detail the sub-safety function deeply with merging features, few works are found to deal with sensor platforms focusing on rear side, and situations of merging lane side with the consideration of relative relation variations with other vehicles and road border markers. However, performance enhancement is needed assuring 100% safety-reliability-optimality and single-objectivity. Also, platforms of on-board sensors including side and rear view are needed to deal with false negative operations and false positive operations. The optimal operation line model of human factors is designed based on time-head-way (reliability), time-to-collision (safety), and combinations of time-head-way and time-to-collision (optimality). The general theory of model predictive control is used to find the target. The model based methodology is applied to solve the human factor model of risk feeling based on only time-head-way and time-to-collision for the human reaction and acceptance metric. Experimental results validated the effectiveness of the proposed approach. The model parameters can be calibrated internationally by tuning the metric of cooperativeness. The target of the predictive judge sub-function is to move the operation point to the specified area. The predictive judge sub-function on high level is decisive for regulation control to move the operation point from difficult areas to the target area in future.
Yu, Kaijiang
Detecting the Driving Intention of the Remote Vehicles Using IMM Estimator2020-01-01104/14/2020
In the development of automated driving vehicle, it is important to detect the driving intentions of the remote vehicles, such as if the remote vehicle on left lane intends to keep driving along the same lane (lane keeping) or change to right lane (change to right lane which results in cut in to host lane), or if the lead vehicle intends to follow the vehicle on the adjacent lane and then change to that adjacent lane. In this paper, we have proposed and implemented a remote vehicle driving intention estimation system which specifically detects the driving intentions of remote vehicles in lateral direction. The estimation FOV covers the three lanes (left, ego, right). The main estimated driving intentions include lane keeping, start lane change to left or right side, arriving to left or ego or right lane, etc. To provide the full FOV range objects data, first, we have proposed and implemented an object tracking algorithm, which tracks the objects in the full three lanes to give us the convenience of analyzing the object motion especially when an object is moving from one lane to another lane (the originally provided object data sets only have the object motion information in the specific lane where the object is located). Then, for all objects, the driving intentions, such as lane keeping, lane change to left/right, are derived based on the object deviation and deviation rate data from the ego lane center. Specifically, we have proposed the kinematic models for vehicle lane keeping and lane changing maneuvers, and used the interacting multiple model algorithm to detect if a vehicle is performing a lane keeping or lane change maneuver by calculating probabilities of the vehicle motion is aligned to the lane keeping model and lane changing model. The remote vehicle driving intention detection is implemented in vehicle level with verifications conducted using actual road data which has shown the promising results.
Chen, Yixin
Understanding How Rain Affects Semantic Segmentation Algorithm Performance2020-01-00924/14/2020
Research interests in autonomous driving have increased significantly in recent years. Several methods are being suggested for performance optimization of autonomous vehicles. However, weather conditions such as rain, snow, and fog may hinder the performance of autonomous algorithms. It is therefore of great importance to study how the performance/efficiency of the underlying scene understanding algorithms vary with such adverse scenarios. Semantic segmentation is one of the most widely used scene-understanding techniques applied to autonomous driving. In this work, we study the performance degradation of several semantic segmentation algorithms caused by rain for off-road driving scenes. Given the limited availability of datasets for real-world off-road driving scenarios that include rain, we utilize two types of synthetic datasets. The first dataset is a pure synthetic rainy dataset which considers the rain droplets on a camera lens, which is suitable for an autonomous vehicle with outside-mounted cameras. This data is generated by the MAVS simulator. In the second dataset, we take good-weather imagery and artificially incorporate rain streaks. By investigating different simulated rain rates, we quantify the performance of such algorithms and witness the severe performance degradation with increasing rain density. We also propose and analyze two methods to obtain the robust performance of segmentation algorithms for both clear and rainy weather.
Sharma, SuvashGoodin, ChrisDoude, MatthewHudson, ChristopherCarruth, DanielTang, BoBall, John
A Novel Velocity Planner for Autonomous Vehicle Considering Human Driver’s Habits2020-01-01334/14/2020
In automatic driving application, the velocity planner can be considered as a key factor to ensure the safety and comfort. One of the most important tasks of the velocity planner is to simulate the velocity characteristics of human drivers. In this paper, two Driver In-the-Loop (DIL) experiments are designed to explain velocity characteristics of human drivers. In the first experiment, static obstacles are placed on both sides of the straight road to shorten the cross range that vehicles can driver across. Moreover, different cross ranges are set to study the influence of the steering wheel error. In the second experiment, velocity characteristics are investigated under the condition of different road widths and curvatures in a U-turn road contour. In both tests, different drivers’ preview behavior is analyzed through the operation of throttle, braking, and steering. From the results we could see the change of vehicle speed depends largely on the traffic environment at the driver’s preview point. On this basis, a novel velocity planner is proposed. Firstly, a target velocity in preview terminal point is calculated. The calculation of the velocity is based on two indicators-the driver’s driving & operating ability, and the degree of visual restriction. The former refers to the ability of the driver to maintain the driveway as well as the control ability of the vehicle stability, and the latter is related to the uncertainty of the environment. Subsequently, the smooth velocity profiles that connect the initial point and the preview terminal point are generated based on the convex optimization. Finally, the simulation results show that this velocity planner possesses good human-like performance, considering the human-vehicle-road coordination. This study is useful to customize velocity planning for autonomous vehicle so as to improve the acceptability of the specific human driver.
Cui, ZongweiGuo, XuexunPei, Xiaofei
Decision Making and Trajectory Planning for Lane Change Control Inspired by Parallel Parking2020-01-01344/14/2020
Lane-changing systems have been developed and applied to improve environmental adaptability of advanced driver assistant system (ADAS) and driver comfort. Lane-changing control consists of three steps: decision making, trajectory planning and trajectory tracking. Current methods are not perfect due to weaknesses such as high computation cost, low robustness to uncertainties, etc. In this paper, a novel lane changing control method is proposed, where lane-changing behavior is analogized to parallel parking behavior. In the perspective of host vehicle with lane-changing intention, the space between vehicles in the target adjacent lane can be regarded as dynamic parking space. A decision making and path planning algorithm of parallel parking is adapted to deal with lane change condition. The adopted algorithm based on rules checks lane-changing feasibility and generates desired path in the moving reference system at the same speed of vehicles in target lane. Compared to algorithm for static parking space, the uncertainty of the space between moving vehicles and host vehicle dynamics raises stricter requirements for algorithms. Works are conducted to deal with dynamically changing scenarios, such as design of safety zone and exit conditions to avoid collision. Simulation under PreScan-Simulink environment shows that the proposed method outperforms in lane change scenarios and achieves strong robustness to inter-vehicle dynamics.
Yu, LiangyaoRu, ZeLu, ZhenghongLiang, GuanqunXiong, CenboLanie, AbiWang, Ruyue
Series Fuzzy PID with Anti-windup Controller for Intelligent Vehicle2020-01-01134/14/2020
A series fuzzy PID controller with anti-windup scope (SFPCA) is proposed in this paper to address saturation nonlinear problem and control disturbance caused by uncertainty of actuator model. In order to achieve novel dynamic and steady-state performance, the fuzzy controller and PID controller are fused into series, which realizes excellent dynamic performance of fast response and low overshoot like pure fuzzy controller at the initial response stage, and the excellent steady-state performance of stable and no static difference like PID control at the later response stage. The Hurwitz low is employed to configure PID parameters and 49 rules are designed for fuzzy controller. Since the input of the actuator could not be infinite, the actuator being saturated for a long time could reduce the stability of system and, even lead to irreversible damage. Moreover, after exiting the saturation state, it is difficult to quickly recover to the fast and stable response state of the original system. Therefore, an anti-windup scope is meticulously developed to limit the system input to a reasonable range under the saturation state, and, in the unsaturated state, the original Fuzzy PID control is restored. In order to verify the performance of the algorithm, four comparison algorithms were adopted, including pure PD, pure PID, pure Fuzzy and series Fuzzy PID controller (SFPC), and two typical commands like step and sine are employed as desired signals. The experimental results show that the SFPCA has more excellent dynamic and steady performance than pure PD, pure PID, pure Fuzzy and series Fuzzy PID controller (SFPC).
Luo, Chao
Challenges in Integrating Cybersecurity into Existing Development Processes2020-01-01444/14/2020
For an established development process and a team accustomed to this process, adding cybersecurity features to the product initially means inconvenience and reduced productivity without perceivable benefits. Adapting development processes to take cybersecurity into account introduces challenges not present in engineering divisions so far. Strategies designed to deal with these challenges differ in the way in which added duties are assigned and cybersecurity topics are integrated into the already existing process steps. Cybersecurity requirements often clash with existing system requirements or established development methods, leading to low acceptance among developers, and introducing the need to have clear policies on how friction between cybersecurity and other fields is handled. A cybersecurity development approach is frequently perceived as introducing impediments, that bear the risk of cybersecurity measures receiving a lower priority to reduce inconvenience. Moreover, this leads to frustration among cybersecurity developers when their proposals are not accepted, and they feel their work is not appreciated. On the other hand, putting too much emphasis on cybersecurity leads to feature creep and makes the development unnecessarily complicated without producing appropriate results. It seems natural to orientate oneself by how safety topics are handled in the development process and adjust this to accommodate cybersecurity. It is, however, not clear in which way these added responsibilities should be assigned, as conflicts of interest occur when a single person must additionally take cybersecurity goals into account, which might be clashing with other project goals this person is responsible for. Ideally, cybersecurity aspects are considered and integrated into development processes not only to fulfill customer and legal requirements, but also to enable developers of functionalities not directly related to cybersecurity to produce better and more robust results as shortcuts are no longer easily possible.
Lenhart, PatricArndt, Paulvon Wedel, JanaBeul, ChristianWeldert, Jan
Items per page:
1 – 50 of 375