Thesis (M.Tech.:Civil Engineering)-Dept of Civil Engineering and Survey, Durban Institute of Technology, 2005 xiv, 134 leaves, Annexures A-C / Road fatalities claim more than one million lives annually worldwide. The emotional, social and economic impact of road traffic fatalities demands urgent attention globally. This epidemic of road traffic fatalities is plaguing everyone, especially the poorer nations. Some countries like Australia and Sweden have been more successful than others in combating this epidemic. South Africa is currently seeking strategies to combat this epidemic because South Africa’s road traffic fatalities have been increasing annually, with a substantial percentage of teenagers and young adults between the ages of 16 and 29 contributing to these statistics. This age group will become or already have become part of the economically active population and concern is mounting as to why this particular age group is vulnerable. This thesis provides an overview of the road safety problem globally, nationally, provincially and locally and also looks at the historical factors that have contributed to this problem. The Victorian model, which has been classified as the ‘world’s best practice’, has been reviewed. A pilot survey was conducted at the Mangosuthu Technikon and the focal survey was conducted at the tertiary institutions in the Durban area. Students from this sector were selected as they fall in the most vulnerable age group and data was collected from them on various aspects of road safety. On analysing the data, various problems were identified, in particular, lack of resources and limited education pertaining to road safety. A range of possible solutions is recommended and the focus areas are the 3E’s namely: education, enforcement and engineering. However, the focal recommendation is on education and looks at the possibility of introducing learner’s licence testing to the grade 12 syllabi.

This thesis presents a 3D reconstruction approach to the detection of static obstacles from a single rear view parking camera. Corner features are tracked to estimate the vehicle’s motion and to perform multiview triangulation in order to reconstruct the scene. We model the camera motion as planar motion and use the knowledge of the camera pose to efficiently solve motion parameters. Based on the observed motion, we selected snapshots from which the scene is reconstructed. These snapshots guarantee a sufficient baseline between the images and result in more robust scene modeling. Multiview triangulation of a feature is performed only if the feature obeys the epipolar constraint. Triangulated features are semantically labelled according to their 3D location. Obstacle features are spatially clustered to reduce false detections. Finally, the distance to the nearest obstacle cluster is reported to the driver.

DUTRA, Teófilo Bezerra. Gradient-Based Steering for Vision-Based Crowd Simulation Algorithms. 2015. 122 f. Tese (Doutorado em Ciência da Computação)-Universidade Federal do Ceará, Fortaleza, 2015. / Submitted by Vitor Campos (vitband@gmail.com) on 2016-09-28T21:38:15Z No. of bitstreams: 1 2015_tese_tfdutra.pdf: 18611468 bytes, checksum: efc5f80ee6b8df68191af2bca394a07a (MD5) / Approved for entry into archive by Jairo Viana (jairo@ufc.br) on 2016-09-28T22:54:58Z (GMT) No. of bitstreams: 1 2015_tese_tfdutra.pdf: 18611468 bytes, checksum: efc5f80ee6b8df68191af2bca394a07a (MD5) / Made available in DSpace on 2016-09-28T22:54:58Z (GMT). No. of bitstreams: 1 2015_tese_tfdutra.pdf: 18611468 bytes, checksum: efc5f80ee6b8df68191af2bca394a07a (MD5) Previous issue date: 2015 / Most recent crowd simulation algorithms equip agents with a synthetic vision component for steering. They offer promising perspectives by more realistically imitating the way humans navigate according to what they perceive of their environment. In this thesis, it is proposed a new perception/motion loop to steer agents along collision free trajectories that significantly improves the quality of vision-based crowd simulators. In contrast with previous solutions - which make agents avoid collisions in a purely reactive way - it is suggested exploring the full range of possible adaptations and to retain the locally optimal one. To this end, it is introduced a cost function, based on perceptual variables, which estimates an agent’s situation considering both the risks of future collision and a desired destination. It is then computed the partial derivatives of that function with respect to all possible motion adaptations. The agent adapts its motion to follow the steepest gradient. This thesis has thus two main contributions: the definition of a general purpose control scheme for steering synthetic vision-based agents; and the proposition of cost functions for evaluating the dangerousness of the current situation. Improvements are demonstrated in several cases. / Alguns dos algoritmos mais recentes para simulação de multidão equipam agentes com um sistema visual sintético para auxiliá-los em sua locomoção. Eles oferecem perspectivas promissoras ao imitarem de forma mais realista a forma como os humanos navegam de acordo com o que eles percebem do seu ambiente. Nesta tese, é proposto um novo laço de percepção/ação para dirigir agentes ao longo de trajetórias livres de colisões que melhoram significativamente a qualidade dos simuladores de multidão baseados em visão. Em contraste com abordagens anteriores - que fazem agentes evitarem colisões de maneira puramente reativa - é sugerida a exploração de toda gama de adaptações possíveis e a retenção da que for ótima localmente. Para isto, é introduzida uma função de custo, baseada em variáveis de percepção, que estima a situação atual do agente considerando tanto os riscos de futuras colisões como o destino desejado. São então computadas as derivadas parciais dessa função com respeito a todas adaptações de movimento possíveis. O agente adapta seu movimento de forma a seguir o gradiente descendente. Esta tese possui assim duas principais contribuições: a definição de um esquema de controle de propósito geral para a orientação de agentes baseados em visão sintética; e a proposição de funções de custo para avaliar o perigo da situação atual. As melhorias obtidas com o modelo são demonstradas em diversos casos.

abstract: There has been a vast increase in applications of Unmanned Aerial Vehicles (UAVs) in civilian domains. To operate in the civilian airspace, a UAV must be able to sense and avoid both static and moving obstacles for flight safety. While indoor and low-altitude environments are mainly occupied by static obstacles, risks in space of higher altitude primarily come from moving obstacles such as other aircraft or flying vehicles in the airspace. Therefore, the ability to avoid moving obstacles becomes a necessity for Unmanned Aerial Vehicles. Towards enabling a UAV to autonomously sense and avoid moving obstacles, this thesis makes the following contributions. Initially, an image-based reactive motion planner is developed for a quadrotor to avoid a fast approaching obstacle. Furthermore, A Dubin’s curve based geometry method is developed as a global path planner for a fixed-wing UAV to avoid collisions with aircraft. The image-based method is unable to produce an optimal path and the geometry method uses a simplified UAV model. To compensate these two disadvantages, a series of algorithms built upon the Closed-Loop Rapid Exploratory Random Tree are developed as global path planners to generate collision avoidance paths in real time. The algorithms are validated in Software-In-the-Loop (SITL) and Hardware-In-the-Loop (HIL) simulations using a fixed-wing UAV model and in real flight experiments using quadrotors. It is observed that the algorithm enables a UAV to avoid moving obstacles approaching to it with different directions and speeds. / Dissertation/Thesis / Doctoral Dissertation Computer Science 2015

For the last few years computer vision due to its low exploitation cost and great capabilities has been experiencing rapid growth. One of the research fields that benefits from it the most is the aircrafts positioning and collision avoidance. Light cameras with low energy consumption are an ideal solution for UAVs (Unmanned Aerial Vehicles) navigation systems. With the new Swedish law – unique to Europe, that allows for civil usage of UAVs that fly on altitudes up to 120 meters, the need for reliable and cheap positioning systems became even more dire. In this thesis two possible solutions for positioning problem and one for collision avoidance were proposed and analyzed. Possibility of tracking the vehicles position both from ground and from air was exploited. Camera setup for successful positioning and collision avoidance systems was defined and preliminary results for of the systems performance were presented. / Vision systems are employed more and more often in navigation of ground and air robots. Their greatest advantages are: low cost compared to other sensors, ability to capture large portion of the environment very quickly on one image frame, and their light weight, which is a great advantage for air drone navigation systems. In the thesis the problem of UAV (Unmanned Aerial Vehicle) is considered. Two different issues are tackled. First is determining the vehicles position using one down-facing or two front-facing cameras, and the other is sparse obstacle detection. Additionally, in the thesis, the camera calibration process and camera set up for navigation is discussed. Error causes and types are analyzed.

Le but de la thèse est d’améliorer les simulations de foule dans des situations complexes. Notamment, de comprendre comment les êtres humains naviguent en évitant les collisions avec plusieurs personnes. Pour cela, des expériences furent menées avec l’aide de la réalité virtuelle avec des participants qui devaient naviguer dans un environnement peuplé de plusieurs individus. Les données récoltées durant ces expériences ont été analysées, notamment les trajectoires et les adaptations effectuées par les participants pour éviter des collisions. Le regard des participants a aussi été étudié pour comprendre comment un être humain choisie avec qui interagir. Ces analyses ont permis de mieux comprendre comment les êtres humains réagissent face à des situations complexes. Enfin, avec l’aide de ces analyses, des algorithmes ont été créés pour reproduire la navigation des êtres humains et simulé des foules avec de nombreux humains virtuelles se comportant comme des vrais êtres humains. / The goal of this thesis has been to improve crowd simulation for complex situations. Especially, we tried to understand how humans combine many interactions, during the avoidance of multiple targets for example, and to model this combination process. Some experiments have been performed using Virtual Reality with participants navigating through a heavy populated environment. Data have been recorded and analyzed, especially the trajectories and adaptations made by participants to avoid collisions. The gaze of the participants has also been studied to better understand how humans select the target to interact with. These analyses have helped us better understand how human interact during complex situations. Using these results, algorithms have been designed to reproduced human like navigation and simulate crowd with many virtual human acting like real one.

Over the past few years automotive and technology companies have made significant advances in what has been traditionally a completely human function: driving. Crash avoidance features such as lane departure warning and forward collision warning are becoming increasingly more common and cheaper to obtain, even on non-luxury vehicles. Technology companies and auto manufacturers have announced plans to have self-driving vehicles ready for public use as early as 2020. The mass adoption of automated vehicles (AVs) could significantly change surface transportation as we know it today. This thesis is intended to provide a technical analysis of the potential impacts of AVs on current light-duty vehicle miles traveled (VMT) and parking decisions, the economic desirability of widespread deployment of partially automated technologies, and methods for existing roadways to transition to connected and automated vehicle (CAV) transportation, so that policymakers can make more informed decisions during the transition to CAVs. This work takes a look at AVs from a point in time where vehicles are equipped with driver assistance systems (Level 1) to a point in time where AVs are driverless (Level 5) and can self-park. The results of this work indicate that the fleet-wide adoption of partially automated crash avoidance technologies could provide net-benefit of about $4 billion at current system effectiveness and could provide an annual net-benefit up to $202 billion if all relevant crashes could be prevented. About 25% of all crashes could be addressed by the crash avoidance technologies examined in this dissertation. Over time, as technologies become more effective and cheaper due to economies of scale, greater benefits than the $4 billion could be realized. As automated technologies become more advanced and widespread, existing roadways will need to be able to accommodate these vehicles. This work investigates the effects of a dedicated truck platoon lane on congestion on the Pennsylvania Turnpike and provides a method for existing roadways and highways to determine viable platoon demonstration sites. The initial results suggest that there are several sections of turnpike that could serve as commercial truck platoon demonstration site while still providing a high LOS to all other vehicles. Once AVs can safely and legally drive unoccupied, vehicles will no longer be limited to their driver’s destination and can search for cheaper parking in more distant parking locations. This work simulates a fleet of privately owned vehicles (POVs) in search of cheaper parking in Seattle, using a rectangular grid throughout the study area. Model results indicate that we are not likely to see significant increase in vehicle miles traveled (VMT) and energy use from cars moving from downtown parking lots to cheaper parking in distance locations but at higher penetration rates, parking lot revenues could likely decline to the point where operating a lot is unsustainable economically, if no parking demand management policies are implemented. Driverless vehicles also promise to increase mobility for those in underserved populations. This work estimates bounds on the potential increases in travel in a fully automated vehicle environment due to an increase in mobility from the non-driving and senior populations and people with travel-restrictive medical conditions. Three demand wedges were established in order to conduct a first-order bounding analysis. The combination of the results from all three demand wedges represents an upper bound of 295 billion miles or a 14% increase in annual light-duty VMT for the US population 19 and older. AV technology holds much promise in providing a more accessible and safe transportation system. This thesis can help policymakers and stakeholders maximize the benefits and minimize the challenges.

Today, there are industries that utilize AGVs to transport goods and materials from one location to another. For smaller scale industries it is costly to have a custom made AGV for their manufacturing unit, so they modify the shape of an AGV to accommodate the necessity of carrying and transporting goods. When the shape of an AGV is modified the built-in sensors will not detect the change in shape of the AGV. Hence, there is a risk that the AGV may collide with objects. Also in some AGVs floor sensors are missing to detect the presence of floor/no floor in front of the AGV, which can be a hazardous situation as there are chances of the AGV falling off from the surface. An example of such an AGV is the Patrolbot which can travel around in an industrial premise wirelessly, but needs addition of such sensors to avoid collisions with the modified structure. A Patrolbot has been used in this thesis work and ultrasonic sensors are utilised for obstacle detection with a modified structure and a built-in laser scanner is studied for mapping purpose. The results of this master thesis was that the ultrasonic sensors were tested under various conditions and results were derived. To obtain the same level of results every time it is required to maintain the conditions on which the ultrasonic sensors rely.

This thesis presents a 3D reconstruction approach to the detection of static obstacles from a single rear view parking camera. Corner features are tracked to estimate the vehicle’s motion and to perform multiview triangulation in order to reconstruct the scene. We model the camera motion as planar motion and use the knowledge of the camera pose to efficiently solve motion parameters. Based on the observed motion, we selected snapshots from which the scene is reconstructed. These snapshots guarantee a sufficient baseline between the images and result in more robust scene modeling. Multiview triangulation of a feature is performed only if the feature obeys the epipolar constraint. Triangulated features are semantically labelled according to their 3D location. Obstacle features are spatially clustered to reduce false detections. Finally, the distance to the nearest obstacle cluster is reported to the driver.

In this thesis, an approach for robot motion control with collision avoidance and human-following is investigated. Using velocity potential fields approach in a modified, quasi-harmonic, solution, the navigation controller is developed. A quasi-harmonic function based controller uses harmonic solutions for collision avoidance and smoothly changes toward a non-harmonic solution which tends toward a zero velocity command only when approaching the goal. After the motion controller was created, human-following strategy was designed to let a non-holonomic robot have the ability to follow a human in an unknown environment with obstacles. The approach is based on velocity potential fields that permit to generate velocity vector commands that drive the robot at a safe distance with regard to the human while avoiding obstacles. The quasi-harmonic approach is investigated analytically using symbolic math solutions of MAPLETM as well as in simulations using MATLABTM. Motion simulations of both holonomic and non-holonomic robot motion illustrate how the proposed approach operates. Experiments are also made with LEGO MINDSTROMS NXT robot to test the algorithm in environment with simple and complex obstacles.