No description available.

With the advent of autonomous vehicles, the issue of safety-evaluationhas become key. ISO26262 recommends using Markov chains. However, in their most common form, Markov chains lack the flexibility required to model non- exponential probability distributions and systems displaying parallelism. In these cases, generalized semi-Markov processes arebetter suited. Though, these are significantly more taxing to analyze mathematically.  This thesis instead explores the option of simulating these systemsdirectly via MATLAB’s Simulink and Stateflow. An example system, here called CASE, currently under study by Scania was used as an example. The results showed that direct simulation is indeed possible, but the computational times are significantly greater than those from standard MATLAB-functions. The method should therefore be employed on parallel systems when results with a high level of fidelity are needed, and alternative methods are not available.

Résumé indisponible. / The VIPAFLEET project aims at developing a framework to manage a fleet of IndividualPublic Autonomous Vehicles (VIPA). We consider a fleet of cars distributed at specifiedstations in an industrial area to supply internal transportation, where the cars can beused in different modes of circulation (tram mode, elevator mode, taxi mode). The goalis to develop and implement suitable algorithms for each mode in order to satisfy all therequests either under an economic point aspect or under a quality of service aspect, thisby varying the studied objective functions.We model the underlying online transportation system as a discrete event basedsystem and propose a corresponding fleet management framework, to handle modes,demands and commands. We consider three modes of circulation, tram, elevator andtaxi mode. We propose for each mode appropriate online algorithms and evaluate theirperformance, both in terms of competitive analysis and practical behavior by computationalresults. We treat in this work, the pickup and delivery problem related to theTram mode and the Elevator mode the pickup and delivery problem with time windowsrelated to the taxi mode by means of flows in time-expanded networks.

abstract: The autonomous vehicle technology has come a long way, but currently, there are no companies that are able to offer fully autonomous ride in any conditions, on any road without any human supervision. These systems should be extensively trained and validated to guarantee safe human transportation. Any small errors in the system functionality may lead to fatal accidents and may endanger human lives. Deep learning methods are widely used for environment perception and prediction of hazardous situations. These techniques require huge amount of training data with both normal and abnormal samples to enable the vehicle to avoid a dangerous situation. The goal of this thesis is to generate simulations from real-world tricky collision scenarios for training and testing autonomous vehicles. Dashcam crash videos from the internet can now be utilized to extract valuable collision data and recreate the crash scenarios in a simulator. The problem of extracting 3D vehicle trajectories from videos recorded by an unknown monocular camera source is solved using a modular approach. The framework is divided into two stages: (a) extracting meaningful adversarial trajectories from short crash videos, and (b) developing methods to automatically process and simulate the vehicle trajectories on a vehicle simulator. / Dissertation/Thesis / Video Demonstration / Masters Thesis Computer Science 2019

<div>Collision avoidance design plays an essential role in autonomous vehicle technology. It's an attractive research area that will need much experimentation in the future. This research area is very important for providing the maximum safety to automated vehicles, which have to be tested several times under different circumstances for safety before use in real life. This thesis proposes a method for designing and presenting a collision avoidance maneuver by using a model predictive controller with a moving obstacle for automated vehicles. It consists of a plant model, an adaptive MPC controller, and a reference trajectory. The proposed strategy applies a dynamic bicycle model as the plant model, adaptive model predictive controller for the lateral control, and a custom reference trajectory for the scenario design. The model was developed using the Model Predictive Control Toolbox and Automated Driving Toolbox in Matlab. Builtin tools available in Matlab/Simulink were used to verify the modeling approach and analyze the performance of the system. The major contribution of this thesis work was implementing a novel dynamic obstacle avoidance control method for automated vehicles. The study used validated parameters obtained from previous research. The novelty of this research was performing the studies using a MPC based controller instead of a sliding mode controller, that was primarily used in other studies. The results obtained from the study are compared with the validated models. The comparisons consisted of the lateral overlap,lateral error, and steering angle simulation results between the models. Additionally,this study also included outcomes for the yaw angle. The comparisons and other outcomes obtained in this study indicated that the developed control model produced reasonably acceptable results and recommendations for future studies.</div>

<div> <div> <div> <p>With the advancement in Unmanned Aerial Vehicles (UAV) technology, UAVs have become accessible to the public. However, recent world events have highlighted that the rapid increase of UAVs is bringing with it a threat to public privacy and security. Thus, it is important to think about how to prevent the threats of UAVs to protect our privacy and safety. This study aims to provide an alternative way to substitute an expensive system by using 2D optical sensors that can be easily utilized by people. One of the main challenges for aerial object recognition with computer vision is discriminating other flying objects from the targets, in the far distance. There are limitation to classify the flying object when it appears as a set of small black pixels on the frame. The movement feature can help the system to extract the discriminative feature, so that the classifier can classify the UAV and other objects, such as a bird. Thus, this study proposes a drone detection system using two elements of information, which are appearance information and motion information to overcome the limitation of a vision based system. </p> </div> </div> </div>

<div><div><div><p>Towing capacity affects a vehicle’s towing ability and it is usually costly to buy or even rent a vehicle that can tow certain amount of weight. A widely swaying towing trailer is one of the main causes for accidents that involves towing trailers. This study propose an affordable automated disconnected towing system (ADTS) that does not require physical connection between leading vehicle and the trailer vehicle by only using a computer vision system. The ADTS contains two main parts: a leading vehicle which can perform lane detection and a trailer vehicle which can automatically follow the leading vehicle by detecting the license plate of the leading vehicle. The trailer vehicle can adjust its speed according to the distance from the leading vehicle.</p></div></div></div>

Cyber-Physical Systems (CPSs) are critical to many aspects of our daily lives. Autonomous cars, life saving medical devices, drones for package delivery, and robots for manufacturing are all prime examples of CPSs. The dual cyber/physical operating nature and highly integrated feedback control loops of CPSs means that they inherit security problems from traditional computing systems (e.g., software vulnerabilities, hardware side-channels) and physical systems (e.g., theft, tampering), while additionally introducing challenges of their own. The challenges to achieving security for CPSs stem not only from the interaction of the cyber and physical domains, but from the additional pressures of resource constraints imposed due to cost, limited energy budgets, and real-time nature of workloads. Due to the tight resource constraints of CPSs, there is often little headroom to devote for security. Thus, there is a need for low overhead deployable solutions to harden resource constrained CPSs. This dissertation shows that security can be effectively integrated into resource constrained cyber-physical system devices by leveraging fundamental physical properties, & tailoring and extending age-old abstractions in computing. To provide context on the state of security for CPSs, this document begins with the development of a unifying framework that can be used to identify threats and opportunities for enforcing security policies while providing a systematic survey of the field. This dissertation characterizes the properties of CPSs and typical components (e.g., sensors, actuators, computing devices) in addition to the software commonly used. We discuss available security primitives and their limitations for both hardware and software. In particular, we focus on software security threats targeting memory safety. The rest of the thesis focuses on the design and implementation of novel, deployable approaches to combat memory safety on resource constrained devices used by CPSs (e.g., 32-bit processors and microcontrollers). We first discuss how cyber-physical system properties such as inertia and feedback can be used to harden software efficiently with minimal modification to both hardware and software. We develop the framework You Only Live Once (YOLO) that proactively resets a device and restores it from a secure verified snapshot. YOLO relies on inertia, to tolerate periods of resets, and on feedback to rebuild state when recovering from a snapshot. YOLO is built upon a theoretical model that is used to determine safe operating parameters to aid a system designer in deployment. We evaluate YOLO in simulation and two real-world CPSs, an engine and drone. Second, we explore how rethinking of core computing concepts can lead to new fundamental abstractions that can efficiently hide performance overheads usually associated with hardening software against memory safety issues. To this end, we present two techniques: (i) The Phantom Address Space (PAS) is a new architectural concept that can be used to improve N-version systems by (almost) eliminating the overheads associated with handling replicated execution. Specifically, PAS can be used to provide an efficient implementation of a diversification concept known as execution path randomization aimed at thwarting code-reuse attacks. The goal of execution path randomization is to frequently switch between two distinct program variants forcing the attacker to gamble on which code to reuse. (ii) Cache Line Formats (Califorms) introduces a novel method to efficiently store memory in caches. Califorms makes the novel insight that dead spaces in program data due to its memory layout can be used to efficiently implement the concept of memory blacklisting, which prohibits a program from accessing certain memory regions based on program semantics. Califorms not onlyconsumes less memory than prior approaches, but can provide byte-granular protection while limiting the scope of its hardware changes to caches. While both PAS and Califorms were originally designed to target resource constrained devices, it's worth noting that they are widely applicable and can efficiently scale up to mobile, desktop, and server class processors. As CPSs continue to proliferate and become integrated in more critical infrastructure, security is an increasing concern. However, security will undoubtedly always play second fiddle to financial concerns that affect business bottom lines. Thus, it is important that there be easily deployable, low-overhead solutions that can scale from the most constrained of devices to more featureful systems for future migration. This dissertation is one step towards the goal of providing inexpensive mechanisms to ensure the security of cyber-physical system software.

The provision of an adequate liability regime for ADAS technologies is an essential prerequisite for its roll out over the coming decade. Facing to the challenge of future highly automated vehicles, this paper proposed a Human-Machine Transition (HMT) approach as a common conceptual framework for considering Human Machine Interaction (HMI), liability and ethical issues in a unified way. The issues that arise are interrogated from a legal perspective, more specifically liability regimes and that of applied ethics. The paper highlights the issue of the handover/takeover. Potential consequences for insurance companies are then identified accordingly, with the aim to progress towards the sustainable deployment of automated vehicles on public roads.

<p>This thesis describes the development of a low-cost, low-power, accurate sensor designed for precise, feedback control of an autonomous vehicle to a hitch. Few studies have been completed on the hitching problem, yet it is an important challenge to be solved for vehicles in the agricultural and transportation industries. Existing sensor solutions are high cost, high power, and require modification to the hitch in order to work. Other potential sensor solutions such as LiDAR and Digital Fringe Projection suffer from these same fundamental problems. </p> <p>The solution that has been developed uses an active stereo vision system, combining classical stereo vision with a laser speckle projection system, which solves the correspondence problem experienced by classic stereo vision sensors. A third camera is added to the sensor for texture mapping. As a whole, the system cost is $188, with a power usage of 2.3 W.</p> <p>To test the system, a model test of the hitching problem was developed using an RC car and a target to represent a hitch. In the application, both the stereo system and the texture camera are used for measurement of the hitch, and a control system is implemented to precisely control the vehicle to the hitch. The system can successfully control the vehicle from within 35⁰ of perpendicular to the hitch, to a final position with an overall standard deviation of 3.0 mm of lateral error and 1.5⁰ of angular error. Ultimately, this is believed to be the first low power, low cost hitching system that does not require modification of the hitch in order to sense it. </p>