Browse Topic: Autonomous vehicles

Items (1,029)
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
SCOPE IS UNAVAILABLE.
AE-8C2 Terminating Devices and Tooling Committee
Deep learning (DL) models have attained state-of-the-art performance in numerous fields. Nevertheless, for certain real-world applications, existing models encounter diverse challenges, ranging from a lack of generability to new data to issues of scalability and overfitting. In this context, integrating information extracted from different modalities holds promise as a potential solution to alleviate these challenges. This paper introduces MAVEN, a multimodal deep-learning framework for long-range atmospheric visibility estimation. Using multimodal deep learning, MAVEN fuses various modalities to estimate long-range atmospheric visibility. These modalities include RGB imagery, Edge Map, Entropy Map, Depth Map, and Normal Surface Map. Results show that in contrast to single-modality RGB, which achieves only 87.92% accuracy, multimodal deep learning models achieve an accuracy of over 96%. This significant improvement highlights the potential of multimodal approaches to enhance the accuracy and reliability of atmospheric visibility estimation, which is crucial for improving safety in applications such as aviation, maritime navigation, and autonomous vehicles. By addressing challenges such as data variability, environmental factors, and the inherent complexity of atmospheric conditions, MAVEN contributes to more reliable and robust visibility estimation systems, thereby enhancing safety and operational efficiency in critical environments.
Khelifi, AmineJohnson, CharlesBouaynaya, NidhalCarannante, GiuseppinaBouhsine, Taha
test
The recent discovery of glacier remains in Noctis Labyrinthus, the "Maze of the Night" near Mars' equator sheds new light on the history of water on Mars, the evolution of the planet’s climate and geology, and the possibility of life. It also opens the possibility for massive amounts of clean glacier ice to be accessed by astronauts at low latitudes on Mars, alleviating the need to operate in more frigid higher latitudes. Further reconnaissance of the site requires a robotic vehicle capable of traversing rough, salt-crusted glacier surfaces and leaping across crevasse fields. To address this need, we propose a conceptual hybrid aerial/ground vehicle, LILI (Long-term Ice-field Levitating Investigator). LILI combines episodic rotary-wing flight with ground mobility as a propeller-driven sled through an arrangement of skis/runners, wheels, and tilting proprotors. A high-level look at the Noctis Labyrinthus "relict glacier" site is presented, along with a notional LILI mission traverse concept designed to ensure critical scientific measurements are captured. The NASA Design and Analysis of Rotorcraft (NDARC) software is utilized to ensure that mission requirements and sizing constraints are met. Furthermore, future work considers guidance, navigation, and control requirements to satisfy mission objectives, and an initial construction for a simplified LILI small-scale prototype.
Schatzman, NatashaYoung, LarryDominguez, MichelleLee, PascalNagami, KeikoCaudle, DavidPichay, Isabelle
Fault detection in autonomous VTOL aircraft is critical because even minor degradations can quickly destabilize multirotor vehicles in safety-critical environments. However, real-flight fault detection remains challenging due to sensor noise, environmental disturbances, and the nonlinear aeromechanics of multirotor platforms. This study proposes a comprehensive machine-learning framework for rotor fault detection, isolation, and severity prediction using real flight data. A convolutional neural network (CNN) architecture is developed to learn spatio-temporal patterns from multivariate flight dynamics, enabling direct inference of both the faulty rotor and its damage level. The framework is first validated using simulated data generated by our in-house flight dynamic model. Next, to verify the framework using real flight data, a hexcopter was designed, fabricated and flight tested for both nominal and faulty cases by introducing controlled blade-tip breakage. The trained model achieves rotor-wise fault classification accuracies above 99% and sample-wise severity estimation accuracy of 96% within a ±1% tolerance in experimental data, demonstrating strong generalization and supporting real-time health monitoring for autonomous VTOL systems.
Sarker, RipponDabaghian, PedramHalder, AtanuGoyal, Raman
A new motion platform facility was developed at the University of Maryland to study autonomous landing of vertical lift aircraft on moving ship decks. It includes a 6-by-6-ft platform of 1.5 ton payload, a Vicon system, and a quadrotor. The platform can generate high-frequency / high-amplitude stochastic motions in all six degrees of freedom, including simpler motions typical of DDG-51 class vessels at high seas. The quadrotor is custom-built in-house around a feature-based vision system for detection, tracking, and pose estimation of a moving deck. In this paper, the vision data is coupled with the flight controls to execute autonomous landings. The paper presents a preliminary assessment of the platform, the quadrotor, and the performance of the closed-loop algorithm using simple tracking and landing experiments.
Basak, KumardipDatta, Anubhav
The Rotor Blown Wing (RBW) is a tailsitter Vertical Takeoff and Landing (VTOL) Unmanned Aerial System (UAS) configuration that leverages cutting-edge autonomous flight controls through Sikorsky's MATRIX™ technology to create a highly capable, efficient, and scalable technology platform. By combining the benefits of fixed- and rotary-wing aircraft, the RBW configuration eliminates the need for traditional UAS launch and recovery infrastructure. This paper describes the RBW-5 prototype, a 100-pound, dual 5-foot diameter proprotor demonstrator, and discusses the comprehensive evaluation of its design and operability through a combination of flight tests, wind tunnel experiments, and computational fluid dynamics (CFD) simulations. The results demonstrate the maturity of the UAS and highlights key accomplishments of the RBW-5 program, including successful autonomous takeoff and landing and transitions between hover and forward flight, the extraction of critical "blown-physics" underlying RBW aerodynamics, and the validation of CFD models against unpowered and powered wind tunnel data.
Regan, MarcKlimchenko, VeraSargent, CalWallace, BrianRivera, AntonioKaye, JordanSatira, JasonBowles, PatrickColeman, Dustin
Electric Vertical Takeoff and Landing (eVTOL) vehicles undergoing advanced air mobility (AAM) operations feature increasingly autonomous systems (IAS) with non-traditional role allocations. Ensuring the safety of these operations and their novel human–machine teaming (HMT) paradigms requires an appropriate body of knowledge created through relevant, reproducible research. In this paper, we briefly examine the meaning of teaming; current regulation, standards, and guidance; and the knowledge required to build resilient HMTs before turning our attention to how this knowledge is being created by recent research and what conclusions or recommendations can be made. We identify the need for further research into the holistic performance of HMTs, the effect of novel allocations of roles between humans and machines, the ability of humans to provide resilience to unforeseen dangers when acting as a part of these teams; and the characteristics required for clear, timely, and accurate communication between the humans and machines. This work is done in the context of eVTOL aircraft with an indirect flight control system (IFCS) undergoing urban air mobility operations.
Neogi, NatashaGraydon, MalloryHolbrook, JonMaddalon, JeffreyMcCormick, Frank
The vertical flight industry is on its way to a transformative era, with autonomous technologies set to alter aerial vehicle operations. While it seems certain that fully autonomous helicopters will eventually be deployed for a variety of missions, some high-stakes situations—like medical evacuations (MEDEVAC)—will for the foreseeable future demand human participation in the form of Emergency Medical Care-giving Crew. This study describes the testbed built to run and investigate hypothetical future situations in which a helicopter is autonomously piloted while a human medic with no aviation training, subjected to aviation and medical emergencies, manages patient care onboard. A total of 22 participants, with emergency medical technician certification, nursing or a medical board certification, were invited to run and evaluate the use of AI pilot (AP) in different scenarios of medical evacuation under the following emergencies: medical, empty fuel tank, pressure sensor miscalibration, and engine failure. A comprehensive evaluation of both objective and subjective performance metrics revealed that novice medical professionals could effectively execute medical evacuation operations in conjunction with an AI pilot, even during unforeseen circumstances. The analysis of response times unveiled distinct perspectives on how medics perceive and manage various emergency situations when an AP functions as a collaborative and effective team member.
Doda, SanyaFeigh, KarenAgbeyibor, RichardCortes, CarmenKolb, JackMagalhaes, Jose
The National Research Council of Canada is conducting a multi-year autonomous flight systems research and technology development project entitled Advanced Autonomy Systems for Challenging ENvironments Development & Demonstration (AASCEND). As part of AASCEND a no-hover landing capability has been developed and demonstrated in a variety of environmental conditions, including in limited degraded visual environment (DVE) operations. This paper discusses the requirements for no-hover landings, their value within an Autonomous Flight System (AFS), and the implementation of this capability in the NRC's AASCEND autonomous flight system. It presents a methodology for identifying a no-hover landing envelope, taking into account the complex maneuvering required. Within that methodology a proposed set of assessment criteria for no-hover landing performance and behaviour is introduced. The paper reports on the results of applying this methodology to the AASCEND no-hover landing algorithm in a simulation study, along with associated Landing Zone Evaluation (LZE) system performance requirements.
Gowanlock, DerekSyed, MustafaMoshchensky, AntonEllis, KrisCarrothers, BryanGubbels, Arthur
The Dragonfly relocatable lander was selected as NASA's New Frontiers mission in 2019 to explore the organic-rich surface of Titan, Saturn's largest moon. The coaxial quadrotor vehicle will fly to multiple geologic sites covering a distance of over 50 miles near the Titan equator. At each site, Dragonfly will sample materials, determine the surface composition, and investigate how far prebiotic chemistry has progressed on Titan. Upon arrival, the lander will enter the Titan atmosphere protected inside an aeroshell, which will descend and decelerate with parachutes. At an altitude of approximately 1 km above the ground, the lander will separate from the backshell and perform a controlled transition to powered flight. Prior to separation from the backshell and after the heatshield has been ejected, the Preparation for Powered Flight (PPF) sequence will be initiated, which ensures the lander is in a safe and stable state for autonomous descent. A critical element of PPF is the de-spin maneuver, where diagonally opposing rotors rotate at maximum speed to reduce any residual angular momentum by creating a yaw moment in lander body axes. This paper presents high-fidelity computational fluid dynamics simulations of the Dragonfly rotorcraft lander during the PPF sequence. Aerodynamic performance predictions are compared with test data from the National Full-Scale Aerodynamics Complex to validate the simulations and build confidence in the PPF simulation results. Blade-resolved simulations capture the unsteady and complex flow behavior in Titan's dense, low-temperature atmospheric conditions during PPF. The results are analyzed, providing insight into aerodynamic performance and the aerodynamic moments critical for mission success.
Ventura Diaz, PatriciaEdquist, KarlYoon, Seokkwan
In this work, a vision-based solution is developed to address the challenge of landing on a ship deck with precision and accuracy. For an autonomous landing, it is important to have a fast and accurate pose estimation system along with a reliable control strategy. This research uses fractal ArUCo markers instead of multiple separate markers to allow smooth pose estimation at different heights. Pose estimates are further improved using an Extended Kalman Filter, and a tracking algorithm then uses these estimates to guide the landing. A four degree-of-freedom (roll, pitch, heave and sway) simulator platform was built and used to validate the algorithm. The accuracy of the vision system is compared against that of a motion capture system. Real-world experiments were performed on different quadrotors to demonstrate tracking and landing on the platform with sway, roll, and pitch motions. The results show that the system is efficient and reliable in achieving safe and successful landings. The proposed landing system is concluded to be applicable for landings on the deck of the ship under sea-state 4.
Venkatesh, K S
Heavy wind and high sea states pose challenges to operating unmanned rotorcraft on-board a naval ship, in particular the recovery phase. A novel autonomous landing strategy for unmanned rotorcraft is proposed and investigated. The new landing strategy makes use of a prediction of the future deck motion based on a sensor on the ship deck. The study is based on a nonlinear simulation environment which includes the dynamics of a 100 kg unmanned helicopter and the dynamics of an ocean-going patrol vessel of the Royal Netherlands Navy. The performance of the autonomous landing strategy is evaluated for a wide variety of environmental conditions (sea state) and operational conditions (ship speed and heading). The results clearly indicate that the environmental conditions have a strong influence on the landing performance in terms of touchdown velocity and landing accuracy. Furthermore, the autonomous landing strategy is effective in reducing the mean and peak value of the touchdown velocity compared to a standard automatic landing strategy. A reduction of as much as 60% in landing impact is observed for the worst case environmental and operational condition considered. The results confirm and underline the potential of the novel landing strategy.
Zilver, Damyvan Rooij, MichelBakker, Richard
ABSTRACT Automatic guided vehicles (AGV) have made big inroads in the automation of assembly plants and warehouse operations. There are thousands of AGV units in operation at OEM supplier and service facilities worldwide in virtually every major manufacturing and distribution sector. Although today’s AGV systems can be reconfigured and adapted to meet changes in operation and need, their adaptability is often limited because of inadequacies in current systems. This paper describes a wireless navigated (WN) omni-directional (OD) autonomous guided vehicle (AGV) that incorporates three technical innovations that address the shortfalls. The AGV features consist of: 1) A newly developed integrated wireless navigation technology to allow rapid rerouting of navigation pathways; 2) Omnidirectional wheels to move independently in different directions; 3) Modular space frame construction to conveniently resize and reshape the AGV platform. It includes an overview of the AGVs technical features and how the flexibility and agility can be adapted to fit military and commercial application. The AGV is being evaluated as a mobile work station platform and a precise material handling robot.
Cheok, Ka CRadovnikovich, MichoFleck, PaulHallenbeck, KevinGrzebyk, SteveVanneste, JerryLudwig, WolfgangGarner, Robert
ABSTRACT For many rotorcraft platforms, incorrect timing of the autorotation flare and deceleration maneuvers may result in significant aircraft damage and injury to the crew, or worse. There is a clear need for new pilot cueing and control augmentation technologies that lead to a higher probability of a successful autorotation landing. This paper describes a recent effort to develop two different Tau (time-to-contact)-based autorotation controllers that can be used to drive visual aids to help guide a pilot to apply the required control inputs to complete a safe autorotative landing. Such controllers may also be useful for fully autonomous autorotation landing for unmanned vehicles.
Rogers, JonathanJump, MichaelEberle, BrianCameron, Neil
Through the development and flight testing of the Canadian Vertical Lift Autonomy Demonstration (CVLAD) Autonomous Flight System (AFS), the NRC has developed technical and operational insight into many high-level concepts of a full-scale supervised-autonomous helicopter, which are thought to be applicable to a wide variety of implementations and approaches. This paper presents two important concepts: "The Contract" and "Levels of Aggression", for which it is expected some aspect of implementation would be required in any supervised autonomous platform and in particular for platforms where the pilot supervising the autonomy remains on-board the aircraft (and thus the AFS provides a 'competent co-pilot' type functionality).
Gowanlock, DerekComeau, PerryCarrothers, BryanGubbels, ArthurEllis, KrisJennings, Sion
In this paper, we develop and validate a 3D feature-based algorithm for tracking stochastic ship-deck motion at high sea states, specifically Sea-State 6 using data from the Navy SCONE dataset. The new vision algorithm was developed from the structure-from-motion technique, which recovers the 3D structure of an object from a series of 2D images, and was validated using a simulated 3D ship-deck attached to a moving Stewart platform. Algorithm performance with different feature detectors and image resolutions was compared. In hand-held tests, the vision algorithm was demonstrated to accurately estimate the pose of a moving ship-deck using a quadrotor. Visually degraded conditions were also evaluated; the algorithm is robust to occlusion and low illumination, but performance reduces in severe glare. The vision algorithm was then validated in a simple free-flight test. All results were compared with Vicon ground-truth data. Additionally, as the 3D algorithm is computationally demanding, we develop and validate a method to improve the computational speed of the vision algorithm.
Britcher, VictoriaDatta, AnubhavChopra, Inderjit
The paper presents a novel strategy for minimum energy consumption in automatic conversion control of tiltrotor eVTOL aircraft, exemplified by the Aston Martin Volante Vision model. We introduce a tilt schedule methodology that strategically balances conversion and reconversion performance with climb, descent, and cruise phases to minimize overall energy expenditure. Our approach accounts for critical factors such as blade loading, operation handling qualities, and passenger ride comfort within a predefined conversion corridor. The optimized trajectories approximate the minimum energy pathway, essential for operational efficiency in urban air mobility. Analytical results demonstrate that our proposed conversion and reconversion phase profiles significantly reduce energy consumption, contributing to the sustainability of tiltrotor flight operations. This research not only enhances understanding of tiltrotor dynamics but also serves as a pivotal step toward achieving globally optimized energy usage, marking a significant advancement in autonomous flight technology for advanced air mobility systems.
Kang, NamukWhidborne, JamesLu, Linghai
This paper presents an overview of the Autonomous Rotorcraft Project (ARP), a collaborative research initiative launched by the US Army and NASA in 2000 aimed at advancing rotorcraft autonomy. ARP has made substantial progress in areas such as real-time reactive obstacle-avoidance, threat- and terrain-aware navigation, identification of safe landing zones, autonomous flight-control, external sling-load operations, scalable autonomy, and pilot-autonomy interfaces. These advancements have undergone extensive validation through both simulation and flight test. This paper chronicles ARP's research evolution, highlighting milestones achieved and remaining challenges.
Whalley, MatthewWaldman, DavidCarr, JamesOtt, LTC (Ret) CarlOgden, LTC WesleyLusardi, JefferyFujizawa, BrianSchulein, GregoryMielcarek, NathanGoerzen, ChadTakahashi, Marc
Fundamental advancements in aircraft design over the past 50 years have enabled a range of Vertical Takeoff and Landing (VTOL) air vehicle configurations and have significantly enhanced aircraft performance, safety, and reliability. This summary paper chronicles the evolution of rotorcraft design from 1974-2024 as presented by the Vertical Flight Society (VFS) Aircraft Design Technical Committee (TC). It is segregated into three key pillars of aircraft design preceded by an aircraft design overview. The three pillars are: processes and tools, technology, and air vehicle configuration. The first pillar on design processes and tools describes advancements in technology, methodologies, and computational capabilities such as the transition from design solely by wind tunnel testing, physical models, and hand calculations to computer aided design/synthesis and simulation in a model-based engineering digital-twin environment. The second pillar on technology focuses on advances in disciplinary technologies and how they have been incorporated into aircraft design. Technologies discussed include composite materials; rotor systems; propulsion systems, ranging from advancements in turbine engines to all- and more-electric propulsion technologies utilizing various energy storage systems to convertible engines; fly-by-wire systems; avionics and cockpit architectures, including digital displays, navigation aids, and communication equipment; and autonomous and semi-autonomous systems. The third pillar on air vehicle configuration focuses on platform architectures. Major architectures discussed are high-speed VTOLs such as the thrust-vectored aircraft, tiltrotors, lift/thrust compounded helicopters and all- and more-electric aircraft.
Strauss, MichaelScott, Mark
ABSTRACT
Yang, WeiminXiang, SenweiWang, TingYe, MinxiangZhang, YifeiXie, Anhuan
ABSTRACT
Alexander, MarcCraig, GregBorshchova, IrynaGowanlock, DerekGubbels, ArthurEllis, ErisJayasiri, AwanthaNaprstek, TomasCarrothers, BryanJennings, SionComeau, Perry
ABSTRACT
Hendrick, ChristopherJaques, EmmaHorn, JosephLangelaan, JackSydney, Anish
López, JacoboAllen, MichaelLópez, David
ABSTRACT
Lampazzi, MargaretPankok(Jr.),  Carl
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
ABSTRACT
Karakalas, AnargyrosLagoudas,  DimitrisFerede, EtanaGandhi,  Farhan
ABSTRACT
Jennings, SionComeau,  PerryGowanlock,  DerekRobazza,  John
ABSTRACT
Shastry, AbhishekChopra,  InderjitDatta,  Anubhav
Impact of Fog Particles on 1.55 μm Automotive LiDAR Sensor Performance: An Experimental Study in an Enclosed Chamber2021-01-00814/6/2021
To achieve full automation in self-driving vehicles, environmental perception sensing accuracy is critically important. However, ambient particles in adverse weather like foggy, rainy, or snowy conditions can significantly scatter the incident laser beam, and therefore contaminate the intensity and accuracy of light detection and ranging (LiDAR) sensors. Especially compared to the rapidity of technology development in self-driving vehicles, there is a significant lack of documented research on LiDAR systems with wavelength longer than 1 μm for application in Advanced Driver-Assistance Systems. In this work, experimental studies were performed with a state-of-the-art 1.55 μm wavelength automotive-grade LiDAR system in a controlled laboratory fog chamber. The goal of the research is to correlate laser attenuation and the optical properties of fog particles. In this work, a thorough multistep procedure for LiDAR data analysis is presented including spatial averaging of the object measurement and characterizing the temperature effect on a LiDAR intensity parameter. Fog particle density is measured by a commercial visibility sensor instrument. Assuming a constant extinction coefficient and backscatter coefficient, a simple analytical model is derived that correlates LiDAR reflectance and extinction coefficient measured by visibility sensor. Results show that the correlation coefficient between LiDAR and visibility sensor data is 0.98 and the R-squared value of linear fitting is 0.96. By comparing the LiDAR original signal and the model, the Root-Mean-Squared Deviation is 0.007, meaning the model performs very well for predicting LiDAR reflectance in the controlled environment. Furthermore, although the returned signal strength is attenuated, the LiDAR can measure the target with a visibility range lower than six meters.
Zhan, LuNorthrop, William F.
Object Detection and Tracking for Autonomous Vehicles in Adverse Weather Conditions2021-01-00794/6/2021
Object detection and tracking is a central aspect of perception for autonomous vehicles. While there has been significant development in this field in recent years, many perception algorithms still struggle to provide reliable information in challenging weather conditions which include night-time, direct sunlight, glare, fog, etc. To achieve full autonomy, there is a need for a robust perception system capable of handling such challenging conditions. In this paper, we attempt to bridge this gap by proposing an algorithm that combines the strength of automotive radars and infra-red thermal cameras. We show that these sensors complement each other well and provide reliable data in poor visibility conditions. We demonstrate the advantages of a thermal camera over a visible-range camera in these situations and employ YOLOv3 for object detection. The proposed system utilizes a modified Track-Oriented Multiple Hypothesis Tracking (MHT) algorithm which uses data from these sensors to keep track of the surrounding vehicles. The modifications in the well-known MHT algorithm were introduced in order to curb the exponential growth of possible hypotheses and consequently reduce the computational time without loss of any critical information. To validate the system, we provide a real-time implementation on an urban dataset collected at the Texas A&M University.
Bhadoriya, Abhay SinghVegamoor, Vamsi KrishnaRathinam, Sivakumar
A Semantic Slam System Based on Visual-Inertial Information and around View Images for Underground Parking Lot2021-01-00784/6/2021
As one of the most challenging driving tasks, parking is a common but particularly troublesome problem in large cities. Recently, an excellent solution-automated valet parking (AVP) has become a hot research topic, which allows the driver to leave the vehicle in a drop-off area, while the vehicle driving into the parking slot by itself. For AVP, the precise localization is an indispensable module. However, the global positioning system (GPS) cannot be used in the underground parking lot and the localization method based on lidar is too expensive. In response to solve this problem, we propose a simultaneous localization and mapping system with the semantic information of parking slots (PS-SLAM), which is based on visual-inertial and around view images. First, the calibration of multi-sensors is conducted to obtain their intrinsic and extrinsic parameters. In this way, the around view image and transformation matrices between sensors can be acquired. Then, the ORB-SLAM3 based on visual-inertial information is used to acquire the pose of the vehicle and sparse point cloud map. Next, the parking slot in the around view image is detected by the deep convolutional neural network (DCNN) model called VPS-Net. Finally, a parking-slot association method is devised to associate the detected parking slots with the point cloud map to generate a semantic map. The field experiments are conducted using a wire control chassis with 4 fisheye cameras, an inertial measurement unit (IMU), and a monocular camera. The results show that the proposed visual semantic SLAM system not only can achieve centimeter-level localization in the indoor parking lot but also generate a semantic map with parking slots.
LI, WeiLi, ChaohuiXiao, DongjieZhou, DongWang, TaoCao, Libo
Adopting Aviation Safety Knowledge into the Discussions of Safe Implementation of Connected and Autonomous Road Vehicles2021-01-00744/6/2021
The development of connected and autonomous vehicles (CAVs) is progressing fast. Yet, safety and standardization-related discussions are limited due to the recent nature of the sector. Despite the effort that is initiated to kick-start the study, awareness among practitioners is still low. Hence, further effort is required to stimulate this discussion. Among the available works on CAV safety, some of them take inspiration from the aviation sector that has strict safety regulations. The underlying reason is the experience that has been gained over the decades. However, the literature still lacks a thorough association between automation in aviation and the CAV from the safety perspective. As such, this paper motivates the adoption of safe-automation knowledge from aviation to facilitate safer CAV systems. The authors briefly elaborate on the widely discussed aviation themes, including autopilot and auto-throttle malfunctions, flight management system, human factors, and suggests how this knowledge can improve the safety of road CAVs use-case. Besides, the differences between the safety consideration in the two fields are also denoted. In summary, the main aim of this paper is to highlight the potential benefits of adopting aviation automation safety knowledge into safe CAV development. With the advances in the CAV, the authors are convinced that this subject could serve software developers and engineers in developing safe and standardized CAV technology.
Abdul Hamid, Umar ZakirMehndiratta, MohitAdali, Erkan
Development of Fault Detection and Emergency Control for Application to Autonomous Vehicle2021-01-00754/6/2021
This paper describes a failsafe system of automated driving vehicles. The failsafe system consists of the following two parts: sliding mode observer-based environment sensor, chassis sensor fault detection, and emergency deceleration control. Two sliding mode observers are designed to reconstruct the fault of acceleration and environment sensor(Lidar) in a longitudinal direction. In the environment sensor's fault detection part, the longitudinal vehicle model receives clearance and relative velocity values. Therefore, failure diagnosis is possible regardless of environmental sensors, such as radar, lidar, and camera. This paper's sensor data is the failure of Delphi's Electronically Scanning Radar (ESR) and Ibeo's LUX Lidar installed in an autonomous vehicle. The emergency deceleration control algorithm employs the sliding mode control with adaptive convergence time. In the event of a failure, it is significant to control the vehicle within a short period safely. The Adaptive convergence time concept proves a mathematical convergence of the vehicle control time after a failure occurs. As soon as the error occurred, the error was proved to always converge to zero within the final time. Thus, the proposed method introduces the concept of convergence time, and mathematically demonstrates that the state reached the reference target within the specified time when a failure occurred. In the emergency control part, two processing unit hardware structures are adopted to comply with SAE International standard J3016 and NHTSA autonomous vehicle safety report standards. The proposed fail-safe detection algorithm is evaluated through vehicle test data, and the fail-safe control algorithm evaluates through computer simulation and vehicle tests.
Jong Min, LeeOh, Kwang SeokSong, Taejun
Techno-Economic Analysis of Fixed-Route Autonomous and Electric Shuttles2021-01-00614/6/2021
This paper takes a realistic approach to develop a techno-economic analysis for fixed-route autonomous shuttles. To develop a model for analysis, the current state of technology was used to approximate three timelines for achieving SAE level 5 capabilities: progressive, realistic, and conservative. Within these timelines, there are four different increments for advancements in the technology laid out as follows: SAE Level 0 - human driver, SAE Level 4 - in-vehicle safety operator, SAE Level 4 - remote safety operator, and SAE Level 5 - no safety operator. These increments in the changes of the technology were chosen based on the trends in the industry. Various shuttle models were used based on different rider quantities and drive-train requirements (electric vs gas) in this analysis. This allows for further understanding of how these deployment plans will vary the cost for shuttles operating in high, mid, and low ridership demand environments. Additional drive-train comparison shows the savings based on the choice of electric vs ICE vehicles. Taking these parameters into consideration, simulations were run for the various vehicle models in the various ridership demand environments to produce the economic costs for each situation. It was found that in 15 years there is an economic savings of 72%, 68%, 43%, and 35% for small, medium, large, and extra-large shuttles, respectively if deployed with the conservative plan of becoming SAE Level 4 with an in-vehicle safety operator in 2 years, SAE Level 4 with a remote safety operator in 4 years, and SAE Level 5 in 8 years.
Goberville, NickZoardar, Md MarsadRojas, JohanBrown, NicolasMotallebiaraghi, FarhangNavarro, AnthonyAsher, Zachary
Fusing Offline and Online Trajectory Optimization Techniques for Goal-to-Goal Navigation of a Scaled Autonomous Vehicle2021-01-00974/6/2021
Enabling self-driving vehicles to efficiently and autonomously navigate through an obstacle-filled environment remains a topic of significant contemporary research interest. Motion-planning frameworks, encapsulating both path- and trajectory-planning, have played a dominant role in realizing the deployment of a “sense-think-act” intelligence for autonomous vehicles. However, verification and validation of such intelligence on actual self-driving autonomous vehicles has been limited. Simulation-based verification and validation has the advantage of permitting diverse scenario-based testing and comprehensive “what-if” analyses - but is ultimately limited by the simulation fidelity and realism. In contrast, testing on full-scale real-world systems is constrained by the usual challenges of time, space, and cost engendered in reproducing diverse scenarios in practice. Further, motion-planning frameworks often engender a mixture of global-planning (typically performed offline) coupled with a sensor-based local-planning (typically done online), which requires both simulation and physical testing. Thus, scaled vehicle experimentation provides researchers with an exciting via-media to evaluate the performance and robustness of motion-planning algorithms on actual physical hardware - especially in real-time sensor-based motion planning settings. In this paper, we analyze a 1/10th scale F1/10 vehicle's performance in simulation and the actual hardware. A global planning algorithm is used to provide the waypoints for a feasible collision-free path between the start and goal configurations in the environment. We explored the deployment of Rapidly exploring Random Tree (RRT) and Rapidly exploring Random Tree* (RRT*). The Time Elastic Band local trajectory planner in ROS is then used for the realization of smooth, feasible paths between the waypoints. A comparison of validation in simulation has been provided with a detailed discussion of the parametric tuning for improving each case's performance.
Joglekar, AjinkyaDeshpande, BhooshanBasuthakur, MugdhaKrovi, Venkat N
No Cost Autonomous Vehicle Advancements in CARLA through ROS2021-01-01064/6/2021
Development of autonomous vehicle technology is expensive and perhaps more complicated than initially thought, as evidenced by the recent rollback of anticipated delivery dates from companies such as Tesla, Waymo, GM, and more. One of the most effective techniques to reduce research and development costs and speed up implementation is rigorous analysis through simulation. In this paper, we present multiple autonomous vehicle perception and control strategies that are rigorously investigated in the user friendly, free, and open-source simulation environment, CARLA. Overall, we successfully formulated potential solutions to the autonomous navigation problem and assessed their advantages and disadvantages in simulation at no cost. First, a lane finding method utilizing polynomial fitting and machine learning is proposed. Then, the waypoint navigation strategy is described, along with route planning. Object detection is then implemented using pre-trained convolutional neural networks. A classic PID control strategy and the Stanley Method were investigated for lateral and longitudinal control of the vehicle. Finally, each of these components are simultaneously applied in the simulation environment using the robot operating system (ROS). As a result, we have achieved successful self-driving simulation. The key takeaway is that the perception and control strategies proposed can be easily transitioned towards physical implementation, through the use of ROS. The overall conclusion is that the CARLA simulation environment is a reliable workbench to test innovative solutions that could become technology enablers for the autonomous vehicle industry.
Prescinotti Vivan, GabrielGoberville, NickAsher, ZacharyBrown, NicolasRojas, Johan
Predicting Desired Temporal Waypoints from Camera and Route Planner Images using End-To-Mid Imitation Learning2021-01-00884/6/2021
This study is focused on exploring the possibilities of using camera and route planner images for autonomous driving in an end-to-mid learning fashion. The overall idea is to clone the humans’ driving behavior, in particular, their use of vision for ‘driving’ and map for ‘navigating’. The notion is that we humans use our vision to ‘drive’ and sometimes, we also use a map such as Google/Apple maps to find direction in order to ‘navigate’. We replicated this notion by using end-to-mid imitation learning. In particular, we imitated human driving behavior by using camera and route planner images for predicting the desired waypoints and by using a dedicated control to follow those predicted waypoints. Besides, this work also places emphasis on using minimal and cheaper sensors such as camera and basic map for autonomous driving rather than expensive sensors such Lidar or HD Maps as we humans do not use such sophisticated sensors for driving. Also, even after decades of research, the reasonable place for ‘mid’ in the End-to-End approach, as well as, the trade-off between data-driven and math-based approach is not fully understood. Therefore, we focused on the end-to-mid learning approach and tried to identify the reasonable place for ‘mid’ in the end-to-end pipeline.
Arul Doss, Aravind ChandradossGuvenc, Levent
Items per page:
1 – 50 of 1029