We address the task of classifying car images at multiple levels of detail, ranging from the top-level car type, down to the specific car make, model, and year. We analyze existing datasets for car ...classification, and identify the CompCars as an excellent starting point for our task. We show that convolutional neural networks achieve an accuracy above 90% on the finest-level classification task. This high performance, however, is scarcely representative of real-world situations, as it is evaluated on a biased training/test split. In this work, we revisit the CompCars dataset by first defining a new training/test split, which better represents real-world scenarios by setting a more realistic baseline at 61% accuracy on the new test set. We also propagate the existing (but limited) type-level annotation to the entire dataset, and we finally provide a car-tight bounding box for each image, automatically defined through an ad hoc car detector. To evaluate this revisited dataset, we design and implement three different approaches to car classification, two of which exploit the hierarchical nature of car annotations. Our experiments show that higher-level classification in terms of car type positively impacts classification at a finer grain, now reaching 70% accuracy. The achieved performance constitutes a baseline benchmark for future research, and our enriched set of annotations is made available for public download.
Full text
Available for:
IZUM, KILJ, NUK, PILJ, PNG, SAZU, UL, UM, UPUK
The intensive motorization growth observed in emerging and developing economies has attracted increased academic attention. However, many existing studies frequently investigate the car ownership ...determinants that are typical of Western countries and use aggregate measures that mask the role of imported used cars. This implies that there is an important research gap concerning the role of the second-hand vehicles as a source of car ownership growth in emerging and developing countries. This paper aims to reveal the dichotomous character of car ownership growth in an emerging economy and identify the determinants of local primary (new cars) and secondary (imported used cars) car markets. Using data from the Polish Central Vehicle Register containing entries for more than 20 million cars registered and applying the spatial regression models, we disclose that in addition to well-known determinants of car ownership growth, such as income, population density, and housing types, there may be other factors specific to emerging economies driving this process. Specifically, we test the influence of geographical distance on the source of the car supply and the number of companies and entrepreneurs importing and repairing used cars. The findings suggest that future investigations of motorization processes concerning developing and emerging economies should consider the scale of second-hand car imports and its impact on car ownership and seek country-specific determinants of the phenomenon.
Full text
Available for:
CEKLJ, IZUM, KILJ, NUK, PILJ, PNG, SAZU, UL, UM, UPUK
This article presents experience curves and cost-benefit analyses for electric and plug-in hybrid cars sold in Germany. We find that between 2010 and 2016, the prices and price differentials relative ...to conventional cars declined at learning rates of 23 ± 2% and 32 ± 2% for electric cars and 6 ± 1% and 37 ± 2% for plug-in hybrids. If trends persist, price beak-even with conventional cars may be reached after another 7 ± 1 million electric cars and 5 ± 1 million plug-in hybrids are produced. The user costs of electric and plug-in hybrid cars relative to their conventional counterparts are declining annually by 14% and 26%. Also the costs for mitigating CO
and air pollutant emissions through the deployment of electrified cars tend to decline. However, at current levels, NO
and particle emissions are still mitigated at lower costs by state-of-the-art after-treatment systems than through the electrification of powertrains. Overall, the observation of robust technological learning suggests policy makers should focus their support on non-cost market barriers for the electrification of road transport, addressing specifically the availability of recharging infrastructure.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
We investigated motivations of potential earlier and later adopters for adopting sustainable innovations. A large questionnaire study revealed that potential earlier adopters of innovative cars ...evaluated the symbolic attributes of electric cars, but not the instrumental and environmental attributes, more favorably than later adopters. Evaluations of these three electric car attributes predicted the adoption likelihood of potential earlier and later adopters in a similar way. However, potential earlier adopters' evaluations of symbolic attributes predicted their interest (but not their intention) in an electric car more strongly when they perceived more instrumental drawbacks. Apparently, instrumental drawbacks, which are typical at the introduction stage, are not only a barrier for earlier adopters; such drawbacks can enhance interest in an electric car because of what the car can say about them. This suggests that symbolic attributes of sustainable innovations should be stressed as this is likely to promote their adoption.
•Do potential earlier and later adopters differ in what drives their adoption?•Evaluations of instrumental, environmental, and symbolic attributes predict adoption.•Potential earlier adopters were more positive about the symbolic attributes.•Instrumental drawbacks can boost potential earlier adopters' symbolic motives for adoption.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
Due to the complexity of the driving environment and the dynamics of the behavior of traffic participants, self-driving in dense traffic flow is very challenging. Traditional methods usually rely on ...predefined rules, which are difficult to adapt to various driving scenarios. Deep reinforcement learning (DRL) shows advantages over rule-based methods in complex self-driving environments, demonstrating the great potential of intelligent decision-making. However, one of the problems of DRL is the inefficiency of exploration; typically, it requires a lot of trial and error to learn the optimal policy, which leads to its slow learning rate and makes it difficult for the agent to learn well-performing decision-making policies in self-driving scenarios. Inspired by the outstanding performance of supervised learning in classification tasks, we propose a self-driving intelligent control method that combines human driving experience and adaptive sampling supervised actor-critic algorithm. Unlike traditional DRL, we modified the learning process of the policy network by combining supervised learning and DRL and adding human driving experience to the learning samples to better guide the self-driving vehicle to learn the optimal policy through human driving experience and real-time human guidance. In addition, in order to make the agent learn more efficiently, we introduced real-time human guidance in its learning process, and an adaptive balanced sampling method was designed for improving the sampling performance. We also designed the reward function in detail for different evaluation indexes such as traffic efficiency, which further guides the agent to learn the self-driving intelligent control policy in a better way. The experimental results show that the method is able to control vehicles in complex traffic environments for self-driving tasks and exhibits better performance than other DRL methods.
► Car use is strongly gendered both in car deficient households and in households with as many cars as drivers. ► Driving is positively affected by household work and employed work responsibilities. ...► The hypothesis of intra-household economic power relations affecting car use is not supported. ► A strong ‘sex’ effect may point towards patriarchy and/or preference.
This paper studies travel mode choice with a focus on car use in car deficient households from a gender perspective. Car deficient households are defined as households with more drivers than cars. We derive some key hypotheses from the literature and use the German Mobility Panel 1994–2008 to simultaneously test some of these hypotheses in a pooled data approach with cluster robust regression techniques. We find support for the social roles hypothesis which claims that mode choice may be impacted by the gendered roles a person takes in a household. Participation in paid work does not systematically affect car use more strongly than participation in unpaid work. Thus, there is no support for the economic power hypothesis which claims that car access is a function of intrahousehold economic power. The strong effect of ’sex’ leads us to conclude that there must be more behind gender differences in mode choice than just social roles. Gender differences in travel mode choice even in households with as many cars as drivers suggest that preferences may be at play. The paper concludes with an outlook on further research.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
Imitation learning techniques aim to mimic human behavior in a given task. An agent (a learning machine) is trained to perform a task from demonstrations by learning a mapping between observations ...and actions. The idea of teaching by imitation has been around for many years; however, the field is gaining attention recently due to advances in computing and sensing as well as rising demand for intelligent applications. The paradigm of learning by imitation is gaining popularity because it facilitates teaching complex tasks with minimal expert knowledge of the tasks. Generic imitation learning methods could potentially reduce the problem of teaching a task to that of providing demonstrations, without the need for explicit programming or designing reward functions specific to the task. Modern sensors are able to collect and transmit high volumes of data rapidly, and processors with high computational power allow fast processing that maps the sensory data to actions in a timely manner. This opens the door for many potential AI applications that require real-time perception and reaction such as humanoid robots, self-driving vehicles, human computer interaction, and computer games, to name a few. However, specialized algorithms are needed to effectively and robustly learn models as learning by imitation poses its own set of challenges. In this article, we survey imitation learning methods and present design options in different steps of the learning process. We introduce a background and motivation for the field as well as highlight challenges specific to the imitation problem. Methods for designing and evaluating imitation learning tasks are categorized and reviewed. Special attention is given to learning methods in robotics and games as these domains are the most popular in the literature and provide a wide array of problems and methodologies. We extensively discuss combining imitation learning approaches using different sources and methods, as well as incorporating other motion learning methods to enhance imitation. We also discuss the potential impact on industry, present major applications, and highlight current and future research directions.
Full text
Available for:
IZUM, KILJ, NUK, PILJ, SAZU, UL, UM, UPUK
CAR-T cell therapy, as a novel immunotherapy approach, has indicated successful results in the treatment of hematological malignancies; however, distinct results have been achieved regarding solid ...tumors. Tumor immunosuppressive microenvironment has been identified as the most critical barrier in CAR-T cell therapy of solid tumors. Developing novel strategies to augment the safety and efficacy of CAR-T cells could be useful to overcome the solid tumor hurdles. Similar to other cancer treatments, CAR-T cell therapy can cause some side effects, which can disturb the healthy tissues. In the current review, we will discuss the practical breakthroughs in CAR-T cell therapy using the multi-targeted and programmable CARs instead of conventional types. These superior types of CAR-T cells have been developed to increase the function and safety of T cells in a controllable manner, which would diminish the incidence of relevant side effects. Moreover, we will describe the capability of these powerful CARs in targeting multiple tumor antigens, redirecting the CAR-T cells to specific target cells, incrementing the safety of CARs, and other advantages that lead to promising outcomes in cancer CAR-T cell therapy.
Full text
Available for:
EMUNI, FIS, FZAB, GEOZS, GIS, IJS, IMTLJ, KILJ, KISLJ, MFDPS, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, SBMB, SBNM, UKNU, UL, UM, UPUK, VKSCE, ZAGLJ
Entering an already crowded and established industry, the Niles Car & Manufacturing Company in Ohio began business with surprising success, producing well over 1,000 electric and steam railway ...cars-cars so durable they rarely needed to be replaced. That durability essentially put the company out of business, and it vanished from the scene as quickly as it had appeared, leaving little behind except its sturdy railway cars. The story of this highly regarded company spans just 16 years, from Niles's incorporation in 1901 to the abandonment of railway car production and sale of the property to a firm that would briefly build engine parts during World War I. Including unpublished photographs and rosters of railway cars produced by the company and still in existence in railroad museums, The Electric Pullman will appeal to railroad enthusiasts everywhere.
The connected and automated vehicle (CAV) is a promising piece of technology, anticipated to enhance the safety and effectiveness of mobility. Advanced sensing technologies and control algorithms, ...working to acquire environmental data, analyze data, and regulate vehicle movements, are key functional components of CAVs. In recent years, the creation of innovative sensing technologies for CAVs has gained substantial attention. CAVs can now interpret sensory data to more accurately detect impediments, track their locations, navigate autonomously in a dynamic environment, and communicate with other nearby vehicles. This has been made possible by advancements in sensing technology. Additionally, by utilizing computer vision and other sensing techniques, the bodily movements, facial expressions, and even mental states of in-cabin persons can be identified.