Self-driving cars used to be a concept of a future society. However, through years of research, testing, and dedication they are becoming a modern day reality. To further expand research and testing capabilities in the field of autonomous vehicles, an Autonomous Vehicle Research Platform (AVRP) can be developed. The purpose of an AVRP is to provide researchers with an autonomous ground vehicle testing platform they can outfit with sensors and equipment to meet their specific research needs. The platform will give researchers the capabilities to test algorithms, new sensors, navigation, new technologies, etc. that they believe would help advance autonomous vehicles. When their testing is done, their equipment can be removed so the next researcher can utilize the platform. The scope of this thesis is to develop the operational specifications for an AVRP that can operate at level 4 autonomy. These specifications include navigation and sensing hardware, such as LIDAR, radar, ultrasonic, cameras, and important specifications that pertain to using each, as well as a review of optimal mounting locations. It will also present benchmarks for computing, design specs for power and communication buses, and modifications for universal mounting racks. / Master of Science / A world with self-driving cars may not be as far as we think. Many ground vehicles now a days already have some sort of driver assist system(s) to aid the driver in everyday driving. Examples of these systems include cruise control that adjusts its speed to leading vehicles, or lane detection with steering assist to help keep the vehicle in its lane when the driver is briefly distracted. These smaller systems are far from allowing the vehicle to drive itself, but they do act as a small stepping stone toward fully autonomous vehicles. To further the research and exploration in the world of autonomous ground vehicles, it can be very beneficial to have a single test vehicle that can meet a variety of research needs. This is where an Autonomous Vehicle Research Platform (AVRP) would come in handy. The main goal behind an AVRP is to give researchers the ability to outfit an autonomous research platform with hardware and testing equipment they deem necessary for their research. When the researcher has completed their testing, they remove their added equipment to restore the platform to its base form for the next researcher to use. The scope of this thesis is to develop the operating specifications for an AVRP. This includes types of sensors for understanding the surrounding environment, and their optimal mounting locations, and hardware for positioning and navigating within that environment. It also discusses power estimation for powering the needed hardware and systems, computing benchmarks from other autonomous research platforms, and a communication structure for the AVRP.

With autonomy becoming a reality for passenger cars, developing an autonomous for tractor-trailers is the next step for driverless roads. Tractor-trailers are heavy, large, and have a pivot joint between the tractor and trailer that makes the movement between the two more complicated. The purpose of the research presented here is to determine the best forward "looking" perception sensor that will see far out enough for the vehicle to stop in time to avoid hitting an object. In order to determine the best sensor, a review of previous sensors and autonomous vehicle sensors will be explored along with the various perception technology. Additionally, a simulation of a tractor-trailer stopping was created to determine the range necessary for a forward perception sensor and when jackknifing may occur. The best brake type for a tractor-trailer will be recommended as well. Finally, the best forward sensor and senor layout for an autonomous tractor-trailer is made based upon the simulation results for the stopping distance of a tractor-trailer. The work, however, is not fully complete. A discussion of the future work and validation of the sensors selected will give future research goals. / Master of Science / With autonomy becoming a reality for passenger cars, developing an autonomous for tractor-trailers is the next step for driverless roads. Tractor-trailers are heavy, large, and have a pivot joint between the tractor and trailer that makes the movement between the two more complicated. The purpose of the research presented here is to determine the best forward “looking” perception sensor that will see far out enough for the vehicle to stop in time to avoid hitting an object. In order to determine the best sensor, a review of previous sensors and autonomous vehicle sensors will be explored along with the various perception technology. Additionally, a simulation of a tractor-trailer stopping was created to determine the range necessary for a forward perception sensor and when jackknifing may occur. The best brake type for a tractor-trailer will be recommended as well. Finally, the best forward sensor and senor layout for an autonomous tractor trailer is made based upon the simulation results for the stopping distance of a tractor-trailer. The work, however, is not fully complete. A discussion of the future work and validation of the sensors selected will give future research goals.

Collision avoidance is treated as a game of two players with opposing desiderata. In the application to automated car-like vehicles, we will use a differential game in order to model and assess a worst-case analysis. The end result will be an almost analytic representation of a boundary between a "safe" set and a "unsafe" set. We will generalize the research in [27] to non-identical players and begin the setup of the boundary construction. Then we will consider the advantages and disadvantages of manipulation of the cost function through the solution and control techniques. In particular, we introduce a possible way to incorporate a secondary objective such as sticking to a straight path. We also look a hybrid technique to reduce steering when the opposing player is out of the reach of the vehicle; i.e., is out of the "unsafe" set and less extreme maneuvers may be desired. We first look at a terminal cost formulation and through retrograde techniques may shape this boundary between the "safe" and "unsafe" set. We would like this research, or part thereof, to be assessed and simulated on a simulation vehicle such as that used in the Flexible Low-cost Automated Scaled Highway (FLASH) at the Virginia Tech Transportation Institute (VTTI). In preparation, we briefly look at the sensor demands from this game-theoretic approach. / Master of Science

Autonomous Vehicles (AVs) should result in tremendous benefits to safe human transportation. Recent reports indicate a global average of 3,287 road crash related fatalities a day with the blame, in most cases, assigned to the human driver. By replacing the main cause, AVs are predicted to significantly reduce road accidents -- some claiming up to a 90% reduction on US roads. However, achieving these numbers is not simple. AVs are expected to assume tasks that human drivers perform both consciously and unconsciously -- in some instances, with Machine Learning. AVs incur new levels of complexity that, if handled incorrectly, can result in failures that cause loss of human life and damage to the environment. Accidents involving SAE Level 2 vehicles have highlighted such failures and demonstrated that AVs have a long way to go. The path towards safe AVs includes system architectures that provide effective failure monitoring, detection and mitigation. These architectures must produce AVs that degrade gracefully and remain sufficiently operational in the presence of failures. We introduce Safe-AV, a fault tolerant safety architecture for AVs that is based on the commonly adopted E-Gas 3 Level Monitoring Concept, the Simplex Architecture and guided by a thorough hazard analysis in the form of Systems-Theoretic Process Analysis (STPA). We commenced the architecture design with a review of some modern AV accidents which helped identify the types of failures AVs can present and acted as a first step to our STPA. The hazard analysis was applied to an initial AV architecture (without safety mechanisms) consisting of components that should be present in a typical AV (based on the literature and our ideas). Our STPA identified the system level accidents, hazards and corresponding loss scenarios that led to well-founded safety requirements which, in turn, evolved the initial architecture into Safe-AV. / Thesis / Master of Applied Science (MASc)

International Telemetering Conference Proceedings / October 25-28, 1999 / Riviera Hotel and Convention Center, Las Vegas, Nevada / In this paper, a devising scheme of Autonomous Vehicle Location and Navigation System is introduced firstly. Then, several key technologies used in the devising scheme are presented, which includes a data fusion method based on extended decentralized kalman filter technology, a map-matching method used to compensate the positioning error, and a digital map data processing method used to realize route planning algorithm. By this time, a sample machine based on the devising scheme introduced in this paper has already been worked out successfully. The availability and the advantages of these technologies have been demonstrated.

With the number of deaths, commute time, and injuries constantly rising due to human driving errors, it’s time for a new transportation system, where humans are no longer involved in driving decisions and vehicles are the only machine that decide the actions of a vehicle. To accomplish a fully autonomous world, it’s important for vehicles to be able to communicate instantly and report their movements in order to reduce accidents. This paper discusses four approaches to vehicle to vehicle communication, as well as the underlying standards and technology that enable vehicles to accomplish communicating.

<p>In the context of highway safety factors, road geometrics and pavement condition are of particular interest to highway managers as they fall within their direct control and therefore can be addressed through highway projects. In spite of the preponderance of econometric modeling in highway safety research, there still remain areas worthy of further investigation. These include 1) the lack of sufficient feedback to roadway preservation engineers regarding the impacts of road-surface condition on safety; 2) the inadequate feedback to roadway designers on optimal lane and shoulder width allocation; 3) the need for higher predictive capability and reliability of models that analyze roadway operations; and 4) the lack of realistic simulations to facilitate reliable safety impact studies regarding autonomous vehicles (AV). In an attempt to contribute to the existing knowledge in this domain and to throw more light on these issues, this dissertation proposes a novel framework for enhanced prediction of highway safety that incorporates machine learning and econometrics with optimization to evaluate and quantify the impacts of safety factors. In the traditional highway operating environment, the proposed framework is expected to help agencies improve their safety analysis. Using an Indiana crash dataset, this dissertation implements the framework, thereby 1) estimating the safety impacts of the road-surface condition with advanced econometric specifications, 2) optimizing space resource allocations across highway cross-sectional elements, and 3) predicting the fatality status of highway segments using machine learning algorithms. In addition, this dissertation discusses the opportunities and the expected safety impacts and benefits of AV in the emerging operating environment. The dissertation also presents a proposed deep learning-based autonomous driving simulation framework that addresses the limitations of AV testing and evaluation on in-service roads and test tracks.</p>

Les constructeurs automobiles fabriquant des systèmes de conduite automatisée ont besoin d’aborder les conséquences que ces systèmes peuvent avoir sur la sécurité routière. Notamment pour l’évaluation des gains de sécurité, la sécurisation des essais et l’analyse des accidents impliquant le véhicule autonome. Cependant, le cadre conceptuel actuel utilisé dans la sécurité routière peut ne pas être adapté pour l’analyse des changements et des nouvelles interactions introduits par l’automatisation du véhicule à travers toutes les échelles du système sociotechnique de transport routier.Le but de la thèse est d’appliquer une approche systémique fondée sur STAMP afin d'étudier les gains attendus du véhicule autonome en termes de sécurité routière, sécuriser les expérimentations et analyser les accidents impliquant ce type de véhicule, à travers toutes les échelles du système sociotechnique de transport routier.Afin de contribuer au calcul des gains du véhicule autonome sur la sécurité routière, la population cible d’un « highway pilot system» a été définie et des questions issue d’une analyse STPA (analyse des risques issue de STAMP) aidant à l’évaluation de l’efficacité du système ont été élaborées.Un cadre de sécurisation des expérimentations couvrant tous les niveaux du système a été mis en place au moyen d’une analyse STPA à deux échelles.Enfin, une méthode d’analyse des accidents impliquant un conducteur automatisé a été créé en intégrant des éléments issus de méthodes d’analyses des accidents de la route existantes et des éléments explicatifs développés spécialement à la méthode CAST (méthode d’analyse des accidents fondée sur STAMP). L’accident impliquant une Tesla en mai 2016 est le cas d’étude de cette nouvelle méthode, CASCAD.En conclusion, ces trois applications ont montré tout le potentiel d’une approche systémique fondée sur STAMP pour offrir un cadre conceptuel adapté à l’évaluation des conséquences sur la sécurité routière de la conduite automatisée. / As automakers develop automated driving systems, they must address the implications of such systems on road safety. Notably for the safety benefit assessment, trial safety and accident analysis. However, the existing conceptual framework in road safety may not be adapted to analyze the changes and new interactions introduced by vehicle automation at all the levels of the road transport sociotechnical system.The main objective of this thesis is to apply a systems theoretic approach based on STAMP to examine the safety benefit assessment, trial safety and accident analysis of automated driving across all the levels of the road transport sociotechnical system.This research first contributes to safety benefit assessment by estimating the target population of a highway pilot system and by generating questions derived from an STPA analysis (hazard analysis based on STAMP) to facilitate the evaluation of the influence of the highway pilot system on road safety.Next, this work establishes a framework to ensure trial safety across the macroscopic and microscopic levels of the vehicle trial system by structuring the outputs of two STPA analyses.Finally, this thesis integrates elements from existing crash analysis methods and newly developed guidance elements into CAST (an accident analysis method based on STAMP) to develop a new method for the accident analysis of crashes involving automated driving called CASCAD. The application of CASCAD was illustrated using the available information of the Tesla crash on May 2016.The three applications of this research show the potential of a STAMP-based approach to provide a suitable conceptual framework for the analysis of the implications of road safety on automated driving.

A proposta deste trabalho é contribuir com a construção de uma plataforma de veículo autônomo para viabilizar as pesquisas na área pelo Departamento de Engenharia de Transportes da USP. Até o momento o departamento dispõe de uma plataforma que, a partir de sua posição conhecida, consegue navegar autonomamente até um ponto de destino utilizando apenas uma solução GNSS, no caso, GPS. Para ampliar a mobilidade da plataforma, está sendo sugerida aqui, a adição de sensores inerciais ao veículo, para que ele consiga obter uma solução de posição mesmo em áreas sem cobertura GNSS. Um Sistema de Navegação Inercial não depende de infraestrutura externa, exceto para inicializar suas variáveis, o que neste caso pode ser feito com auxílio de um receptor GPS. Sensores inerciais de alto desempenho são caros, tem alta complexidade mecânica e em geral são de grande porte. A alternativa é o uso de sensores do tipo MEMS que são pequenos, fáceis de serem manipulados e apresentam baixo consumo de energia. A contrapartida é que a solução é mais susceptível a ruído do que seus pares que custam na faixa de centena de milhões de dólares. / The proposal of this paper is to build an autonomous vehicle platform to enable the researches in this area by the Transport Engineering Department of the USP. Until now the Department has a platform that, once its initial position is known, it can navigate autonomously to a destination point using only the GNSS, in this case, GPS. To expand the mobility resources of the platform, it is being suggested here the addition of inertial sensors to the vehicle, enabling it to acquire a position solution even in areas where there is no coverage of the GNSS. An Inertial Navigation System does not depend on an external infra-structure, with the exception on the initial setup, where the GPS can be used to provide this kind of initialization. High performance inertial sensors are expensive, have high mechanical complexity and in general are big. The alternative is the usage of MEMS sensors, which are small, easy to handle and has low power consumption. In the opposite side this solution is more susceptible to noises in comparison to those High performance sensors that cost hundreds of thousands of dollars.

O projeto de mestrado, denominado de forma abreviada como GPUServices, se insere no contexto da pesquisa e do desenvolvimento de métodos de processamento de dados de sensores tridimensionais aplicados a robótica móvel. Tais métodos serão chamados de serviços neste projeto e incluem algoritmos de pré-processamento de nuvens de pontos 3D com segmentação dos dados, a separação e identificação de zonas planares (chão, vias), e detecção de elementos de interesse (bordas, obstáculos). Devido à grande quantidade de dados a serem tratados em um curto espaço de tempo, esses serviços utilizam processamento paralelo por GPU para realizar o processamento parcial ou completo destes dados. A área de aplicação em foco neste projeto visa prover serviços para um sistema ADAS: veículos autônomos e inteligentes, forçando-os a se aproximarem de um sistema de processamento em tempo real devido ao contexto de direção autônoma. Os serviços são divididos em etapas de acordo com a metodologia do projeto, mas sempre buscando a aceleração com o uso de paralelismo inerente: O pré-projeto consiste de organizar um ambiente que seja capaz de coordenar todas as tecnologias utilizadas e que explore o paralelismo; O primeiro serviço tem a responsabilidade de extrair inteligentemente os dados do sensor que foi usado pelo projeto (Sensor laser Velodyne de múltiplos feixes), que se mostra necessário devido à diversos erros de leitura e ao formato de recebimento, fornecendo os dados em uma estrutura matricial; O segundo serviço em cooperação com o anterior corrige a desestabilidade espacial do sensor devido à base de fixação não estar perfeitamente paralela ao chão e devido aos amortecimentos do veículo; O terceiro serviço separa as zonas semânticas do ambiente, como plano do chão, regiões abaixo e acima do chão; O quarto serviço, similar ao anterior, realiza uma pré-segmentação das guias da rua; O quinto serviço realiza uma segmentação de objetos do ambiente, separando-os em blobs; E o sexto serviço utiliza de todos os anteriores para a detecção e segmentação das guias da rua. Os dados recebidos pelo sensor são na forma de uma nuvem de pontos 3D com grande potencial de exploração do paralelismo baseado na localidade das informações. Porém, sua grande dificuldade é a grande taxa de dados recebidos do sensor (em torno de 700.000 pontos/seg.), sendo esta a motivação deste projeto: usar todo o potencial do sensor de forma eficiente ao usar o paralelismo de programação GPU, disponibilizando assim ao usuário serviços de tratamento destes dados. / The master\'s project, abbreviated hence forth as GPUServices, fits in the context of research and development of three-dimensional sensor data processing methods applied to mobile robotics. Such methods will be called services in this project, which include a 3D point cloud preprocessing algorithms with data segmentation, separation and identification of planar areas (ground track), and also detecting elements of interest (borders, barriers). Due to the large amount of data to be processed in a short time, these services should use parallel processing, using the GPU to perform partial or complete processing of these data. The application area in focus in this project aims to provide services for an ADAS system: autonomous and intelligent vehicles, forcing them to get close to a real-time processing system due to the autonomous direction of context.The services are divided into stages according to the project methodology, but always striving for acceleration using inherent parallelism: The pre-project consists of organizing an environment for development that is able to coordinate all used technologies, to exploit parallelism and to be integrated to the system already used by the autonomous car; The first service has a responsibility to intelligently extract sensor data that will be used by the project (Laser sensor Velodyne multi-beam), it appears necessary because of the many reading errors and the receiving data format, hence providing data in a matrix structure; The second service, in cooperation with the above, corrects the spatial destabilization due to the sensor fixing base not perfectly parallel to the ground and due to the damping of the vehicle; The third service separates the environment into semantics areas such as ground plane and regions below and above the ground; The fourth service, similar to the above, performs a pre-segmentation of street cruds; The fifth service performs an environmental objects segmentation, separating them into blobs; The sixth service uses all prior to detection and segmentation of street guides.The received sensor data is structured in the form of a cloud of points. They allow processing with great potential for exploitation of parallelism based on the location of the information. However, its major difficulty is the high rate of data received from the sensor (around 700,000 points/sec), and this gives the motivation of this project: to use the full potential of sensor to efficiently use the parallelism of GPU programming, therefore providing data processing services to the user, providing services that helps and make the implementation of ADAS systems easier and/or faster.