Autonomous Driving Systems (ADSs) have seen rapid progress in recent years. To ensure the safety and reliability of these systems, extensive testing is being conducted before their future mass deployment. One approach is to test ADSs directly on the road, but it is incredibly costly to cover all rare corner cases. Thus, a popular complementary approach is to evaluate an ADS’s performance in a simulator. Such method is called simulation based testing. However, randomly testing ADSs in simulation is still not efficient enough and the testing results might not transfer to the real-world. This dissertation underscores that the cornerstone of efficient simulation testing lies in crafting optimal testing scenarios. We delineate several pivotal properties for these scenarios: they should induce ADS misbehavior, exhibit diversity, manifest realism, and adhere to user specified rules (e.g., following traffic rules). Subsequent to this identification, our research delves into methodologies to enhance one or more of these properties of the generated scenarios. Specifically, we embark on two distinct lines of approach. First, we develop advanced search strategies to unearth diverse scenarios that provoke ADS to misbehave. Second, we harness the potential of deep generative models to produce scenarios that are both realistic and in compliance with user specified rules. Because of the need for efficiently testing end-to-end behaviors of ADSs against complex, real-world corner cases, we propose AutoFuzz, a novel fuzz testing technique, which can leverage widely-used driving simulators’ API grammars to generate complex driving scenarios. In order to find misbehavior-inducing scenarios, which are very rare, we propose a learning based search method to optimize AutoFuzz. In particular, our method trains a neural network to select and mutate scenarios sampled from an evolutionary search method. AutoFuzz shows promises in efficiently identifying traffic violations for the given ADSs under test. Although AutoFuzz is good at finding violations, as a black-box method, it is agnostic of the cause of the violations. In the second project, we focus on finding violations caused by the failure of fusion component, which fuses the inputs of multiple sensors and provides the ADS a more reliable understanding of the surroundings. In particular, we identify that the fusion component of an industry-grade ADAS can fail to trust the more reliable input sensor and thus lead to a collision. We define misbehavior caused by such a failure as "fusion error". In order to efficiently find fusion errors, we propose a fuzzing framework, named FusED, that uses a novel evolutionary-based search method with objective promoting fusion output to deviate from sensor input. We show that FusED can efficiently reveal fusion errors for an industry-grade ADAS. One issue with the generated scenarios by AutoFuzz or FusED (or any other search based methods) is that all the NPC vehicles are controlled by some low-level controllers, whose behaviors are different from human drivers. This poses a difficulty in transferring the found violations into real world. Some recent work tries to address this problem by using deep generative models. However, the scenarios cannot be easily controlled which is not desirable for users to customize the testing scenarios. As both realism and controllability of the generated traffic are desirable, we propose a novel method called Controllable Traffic Generation (CTG) that achieves both properties simultaneously. In order to preserve realism, we propose a conditional, dynamic enforced diffusion model. In order to satisfy controllability, we propose using a kind of "traffic language" called Signal Temporal Logic (STL) to specify what we want in traffic scenarios (e.g., following road rules). We then leverage STL to guide the conditional diffusion model for generating realistic and controllable traffic. Although CTG can generate realistic and controllable traffic, it still requires domain expertise to specify the STL based loss function. Besides, it models traffic participants independently, resulting in sub-optimal agents interaction modeling. In order to address these issues, we developed CTG++ which enables a user to use language to generate realistic traffic scenario. In particular, we proposed to use GPT4 to translate a command in natural language into a loss function in code. We then use the loss function to guide a scene-level diffusion model, which considers all the vehicles jointly, to generate traffic satisfying the command. We have found that CTG++ can generate query (in natural language)-compliant and realistic traffic simulation. In summary, our four projects discussed in this thesis have solved important problems in efficiently testing ADSs and have had significant influence in the advancement of ADS. Besides, the models and empirical studies we performed can be applicable to other testing and behavior generation problems, such as general ML-based software testing, and multi-agent behavior planning and prediction. I hope this thesis can serve as an inspiration to anyone who is interested in the exciting field of ADS testing and development, and contribute to the realization of the full automation of driving.

published_or_final_version / Transport Studies / Master / Master of Arts

This work deals with the analysis of traffic car-accidents in the Czech Republic, Sweden and Denmark. Excepting the analysis, it contains basic information and evaluation. Then in the final part is also given a proposal of measures to reduce car-accidents in the Czech Republic.

The master‘s thesis deals with accidents in road transport in the Czech Republic, Norway and Finland. It includes analysis and comparison of accident rates in those states, an overview of accidents in recent years, legislative requirements and trends in this area and then based on analysis of current proposals includes measures to increase safety in road transport.

The South African taxi industry has experienced large-scale growth in the nineteen nineties and increased its share of the passenger transport market to a majority. However, it tends to suffer from various problems and has been subject to interventions from various quarters, the most recent being the Central Government's recapitalization scheme. The aim of this research was to investigate the perceptions of taxi-owners vis-a-vis the Government's Recapitalization Scheme. More specifically, the objectives were: • To critically appraise the government's recapitalization scheme with a view to assessing its impact on the taxi industry • To establish the economic problems that are faced by the taxi industry in achieving development, thereby empowering the industry and • To identify the capacity-building needs of the taxi-owners. Consequently, while recognizing the notion that 'poor understanding of taxi-owners towards the government's recapitalization scheme has led to resistance on their side'; this research maintains that the taxi-owners' main expectation of the government's recapitalization scheme relates to black economic empowerment. Hence, the view that the government should assist the taxi industry to face the current development challenges. Using both qualitative and quantitative methods, the study focused on short distance taxi-owners under Umsunduzi one municipality in Pietermaritzburg. The main focus of this research was on their perceptions towards the proposed government's recapitalization scheme. Of the 137 taxi-owners, representing 21 Associations, in Pietermaritzburg, the majority of the respondents have positive perceptions towards the scheme. One other concern raised was that the proposed vehicles are too big and that the scheme would impact negatively on some owners who operate on routes where there is only a small number of commuters, as a result the new vehicle will take time to fill with the required number of passengers. The other concern was that some of the drivers and conductors, who are sometimes bread-winners in their homes, were going to lose their jobs. However, they felt that to run their businesses effectively, they would benefit from the government assistance in terms of business skills, education and training. The taxi-owners interviewed, felt that the government is taking too long to implement the scheme and they suggested that the government speed up the process. In addition to the recapitalization scheme, the taxi-owners felt that the government should provide the infrastructure for the scheme, for example taxi ranks with facilities. The taxi-owners believe that conflict within the industry is a threat to their business and they felt that the government should assist them in solving this problem. Findings show that taxi-owners in Pietermaritzburg support the proposed recapitalization scheme. However, some of the taxi-owners are still not clear about the whole process of the recapitalization scheme. The taxi-owners suggested therefore, that the government should communicate more with the taxi-owners in order to implement the scheme successfully. Furthermore, government intervention is required so as to capacitate the taxi-owners by providing them with relevant skills. Infrastructure improvement is required so as to make the taxi business viable. / Thesis (M.A.)-University of Natal, Pietermaritzburg, 2003.

MPM / Oliver Tambo Institute of Governance and Policy Studies / The study investigates the management of traffic law in Limpopo Province with specific reference to Vhembe District and explore possible solutions to the challenges facing Vhembe District in the management of traffic law. Road transport safety, particularly enforcement of traffic laws is challenging globally, especially in developing countries, where it affects both road users and governments. Due to a number of reasons the subculture of traffic and other law enforcement agents is not always viewed in a favorable light. Media reports and newspaper articles give evidence of a total disregard for law enforcement and lack of respect for law enforcement officials. The study used both quantitative and qualitative techniques for data collection and analysis. The Quantitative techniques were mostly used in that they provided the researcher with an understanding of experiences and challenges facing traffic officers in the management of traffic law in Limpopo Province with specific reference to Vhembe District. A simple random and purposive sample was used for selection of a sample population. A purposive sampling technique will be employed in selection study participants of qualitative method. A sample of 50 traffic officers was randomly selected for the survey and 10 traffic chiefs and principals were purposively selected for the interviews. The study found that of lack of good managerial skills, lack of motivation to work, shortage of staff, lack of modern equipment and lack of training as challenges. The treatment of traffic offenses as “petty” by the Department of Justice and in particular magistrates was also identified as a major challenge. The study recommends that the Department of Transport Management should provide traffic police officials with sufficient resources and equipment at the traffic police stations in order to perform optimally. / NRF