Search

Human pose, serving as a robust appearance-invariant mid-level feature, has proven to be effective and efficient for human action recognition and intention estimation. Pose features also have a great potential to improve trajectory prediction for the Vulnerable Road User (VRU) in ADAS or automated driving applications. However, the lack of highly diverse and large VRU pose datasets makes a transfer and application to the VRU rather difficult. This paper introduces the Tsinghua-Daimler Urban Pose dataset (TDUP), a large-scale 2D VRU pose image dataset collected in Chinese urban traffic environments from on-board a moving vehicle. The TDUP dataset contains 21k images with more than 90k high-quality, manually labeled VRU bounding boxes with pose keypoint annotations and additional tags. We optimize four state-of-the-art deep learning approaches (AlphaPose, Mask R-CNN, Pose-SSD and PitPaf) to serve as baselines for the new pose estimation benchmark. We further analyze the effect of using large pre-training datasets and different data proportions as well as optional labeled information during training. Our new benchmark is expected to lay the foundation for further VRU pose studies and to empower the development of accurate VRU trajectory prediction methods in complex urban traffic scenes. The dataset (including an evaluation server) is available on www.urbanpose-dataset.com for non-commercial scientific use.

In this study, Wireless Optical Communication LOS model for underground trains to station based on FSO is designed and simulated to achieve optimum performance and reduced bit error ratio. The FSO medium is highly dependent on Air Quality and Weather conditions bit error ratio vary depending on the conditions. The aim of this design is to achieve a lower Bit error ratio and to increase the range for the communication. The simulation is designed to improve the quality of communication through FSO channel by reducing the bit error ratio through filtering techniques. The results of this simulation will help future research and implementations of the FSO based wireless optical communication for underground trains. This simulation is based on V2I architecture where the communication is based on FSO technology.

The objective of this paper is to develop a computationally efficient and generalised algorithm for the load flow calculation in an active distribution network. The available techniques for the active distribution load flow calculation have limited scope of application and, sometimes, suffer from computational complexity. The applicability of an existing technique remains limited either by DG control modes or by transformer configurations. In this paper, the load flow calculation is carried out by using the concept of Gauss-Zbus iterations, wherein the DG buses are modeled via the technique of power/current compensation. The specific distinctness of the proposed Gauss-Zbus formulation lies in overcoming the limitations imposed by DG control modes for the chosen DG bus modelling as well as in having optimized computational performance. The entire load flow calculation is carried out in the symmetrical component domain by decoupling all the sequence networks. Furthermore, a generalised network modelling is carried out to define decoupled and tap-invariant sequence networks along with maintaining the integrity of the zero sequence network under any transformer configurations.The computational efficiency and accuracy of the methodology proposed are verified through extensive case studies.

The objective of this study is to identify and eliminate unnecessary iteration loops in the load flow analysis of an active distribution network so as to improve its overall computational efficiency. The number of iteration loops is minimised through the integrated modelling of a distributed generator (DG) and the associated coupling transformer. The DG bus is not preserved in the load flow calculation and the aforementioned DG-transformer assembly is represented in the form of a voltage-dependent negative load at the point of connection to the main distribution network. Thus, the iteration stage that is involved in indirectly preserving the DG in the form of a voltage source or negative constant power load can be eliminated. This, in turn, eliminates the need for multiple rounds of forward-backward sweeps (FBS) iterations to determine the bus voltages. The power characteristics of the DG-transformer assembly are thoroughly investigated through a carefully performed case study so as to assess the general convergence performance of the proposed load flow algorithm Furthermore, extensive comparative studies are carried out to verify the computational efficiency attained via the proposed DG modelling in the load flow analysis of an active distribution network.

Priority Preemptive scheduling algorithm is a popular among various other algorithms for scheduling CPU, however it leads to the problem of starvation which happens when processes with lower priority are not given any chance of CPU utilization due to continuous CPU usage by processes with higher priorities. To eradicate this problem here, we have proposed a new algorithm for scheduling CPU which we call it as Modified Priority Preemptive Scheduling Algorithm based on a new approach where we are executing priority Preemptive scheduling in a round robin fashion taking the time quanta equal to the shortest burst time of all the available processes. On analyzing the results, it is observed that modified Preemptive algorithm not only solves the problem of starvation but also enhances the performance of regular Priority Preemptive algorithm by lowering the mean turnaround time and mean waiting time.

The smart grid is widely considered as the advanced version of the power grid. As an essential characteristic of the smart grid, demand response can reschedule the user’s energy consumption to reduce the operating expense from expensive generators, and further to defer the capacity addition in the long run. Demand Response functions as a vital controller for the present electrical framework. From outsider see the interest reaction can be viewed as two connections, one related to customer viewpoint and the other is pertinent to energy supplier or market administrators. But both links are driven by the application of different computational techniques. Throughout the smart grid, the advanced computational technique has been properly utilized to enhance the outcome and capability of learning. The salient feature of this paper is to criticize the applicability of computational techniques for demand response. The paper explores major aspects such as DR issues, different computational techniques, and future extensions of demand response.

Distribution network advancement is the most trending topic in the power system. From reserve capacity reduction by distributed generator (DG) integration to consumer active participation for load management, all the significant changes are occurring in the distribution network. This paper proposes a consumer participation-based load management approach in presence of DGs. Distributed generators such as Photovoltaic (PV) and wind turbines are added to the distribution feeder in constant power voltage mode. Load management is proposed by providing monetary benefits to the consumer for their active participation. Excessive load shifting and formulation of the new peak is prevented by constraints such as thermal, voltage, active and reactive power limit of feeder. Special attention is paid to DGs reactive power supply. The performance of the proposed load management scheme is verified on 15-bus feeder system.

Distribution networks transfer the electrical power the last few miles from transmission to local load centers. These were at first planned as passive systems providing the one directional connection between transmission networks and power end-users. The introduction of Distributed generations (DGs) proposes energy harvesting systems with more flexibility and bidirectional power flows. DGs are intended for generators that are usually employed at a small scale. These generators produce electrical power at places nearer to a consumer or connected to a distribution network. These include both conventional and non-conventional energy resources. In unconnected areas, DGs provide monetary benefits because their installation disposes of the requirement for costly development of distribution and transmission lines. This paper provides a review on significant DG benefits, challenges, applications, drawbacks, and different approaches for DG penetration and gives an insight to power system researchers and engineers for optimal DG planning.

The power system operation needs to keep the harmony among the power generation and demand, failing which will leads to sudden rise and fall of frequency as well as transient tie-line power. AGC is one of the significant control phenomenon in power system, for control of frequency and power in an interconnected system. This paper presents the details of issues related with specific controlling techniques used in AGC at the past and with what control we can rectify those issues in future. This paper also presents a brief review of the controlling methods that has been exercised to automatic generation control (AGC) from the primeval time to the modern day. Review related to literature work of AGC in deregulated environment has been presented.

This paper presents the design, application and analysis of low order modified internal model controller (M-IMC) for automatic voltage regulator (AVR). The drawback of traditional IMC (T-IMC) is the poor input and output disturbance rejection performances that play the key role in AVR for industrial applications. The performance of AVR with M-IMC is compared to T-IMC and ZN tuned PID controller. To the design of low order IMCs, the Routh approximation, improved pole clustering, and truncation methods are applied for the reduction of order of AVR system. The low order controllers are reduces the complexity and computational time of overall system.

This paper is to discuss about the design a fuel car battery charging system using a foot-pump, that charges the battery enough to start the vehicle. A conventional consumer model automobile has a battery capacity of 50Ah to 60Ah that generally lasts for several years. The starting current required for starter motor of such automobiles ranges from 200A to 500A, which is supplied by the battery and the starter circuit. In scenario of failure of the battery to provide sufficient charge, the above-mentioned device would be used. The project is focused elimination of automotive battery problems, which directly or indirectly revolves around the inability of the battery to produce enough charge to start the vehicle. The occurrence of such situations is generally high in regions with extreme temperature, since a battery failure can be caused by both high and low temperatures. The effectiveness as well as demand for such device will be significantly more in cold weathers (especially in areas with perpetual cold climate). The project also rules out the necessity of jump starting the vehicle with the help of another vehicle's battery. Since the uncertainty of availability of another vehicle for jump starting is one of the major causes of being stranded, the device will prove to be of much help in such cases.

So far we have not extended the core part of Visual Studio that is used most often by the developers – the code editor. The Visual Studio editor offers a rich set of extension points, and most of its parts are extensible via the Managed Extensibility Framework (MEF). In fact, there are so many extensibility points exposed by the editor that covering each of these in details would require a book in itself. Fortunately for us, Microsoft has already documented these extensibility points and features in great depth. In this chapter, we will discuss the extensibility points of the Visual Studio editor and develop extensions for code analysis, fixes, and refactoring.

Thus far we have developed and discussed extensions on custom commands, tool windows, infobars, code generation, code editor, code analysis, and refactoring. However, as a developer, we spend most of our time coding, developing, and debugging, so in this chapter, we will look at ways by which we can enhance our productivity while coding and debugging. We will begin with a discussion on code snippets, then move on to project and item templates. Next we will have a look at adding a connected service provider. Finally, will discuss a generic approach to debug an extension. Let’s get things underway!

In the last few chapters, we have gone through a lot of theories, concepts, and fundamentals of building Visual Studio extensions. Now it’s time to put this theory to practice. In this chapter, we will develop a couple of meaningful and useful, real-world Visual Studio extensions in a step-by-step manner. In the process, we will also learn key fundamentals, concepts, and tricks to develop a Visual Studio extension that can be shipped to the Visual Studio Marketplace or shared with a wider audience.

So far we have only discussed how to develop Visual Studio extensions. Equally important is to know and learn how to ship the extension to the Visual Studio Marketplace so that other Visual Studio users can also use the developed extension. In this chapter, we will learn how we can publish our extension to the Visual Studio Marketplace to make it available for the wider audience. With the DevOps getting more traction than ever before, we will see how we can have automated build setup for the extension projects using the Azure DevOps build pipeline. We will then see how we can leverage Azure DevOps (ADO) to automate the process of publishing an extension to the marketplace. Apart from the Visual Studio Marketplace, there may be scenarios to host extensions in a private extension gallery to share extensions with only a subset of users. So, we will see how we can develop a private extension gallery and host extensions. By the end of this chapter, readers should be comfortable in

This chapter builds on the last chapter and continues our odyssey of developing Visual Studio extensions. In this chapter, we will develop a new set of extensions for displaying an information bar and custom code generation. While developing these extensions, we will also advance our knowledge and learn new aspects and concepts of Visual Studio.

In this chapter, we will learn about the extensibility model of Visual Studio. To do so, we will look at the Visual Studio user interface and its software development kit (SDK) and check out the extensibility points available to us. We will then develop a basic extension for a custom command and see how it can be integrated in a code window, tool menu, and solution explorer. We conclude this chapter by discussing some potent questions around command extensibility. By the end of this chapter, we will understand the fundamentals of Visual Studio Extensibility and have developed a basic extension giving us a solid foundation to create the real-world extensions for our beloved IDE.

In a typical wireless sensor network (WSN), a central controller collects information from multiple sensors which are battery operated. The challenging problem is how to minimize the number of re-transmissions or increase the sleep time of sensors without compromising on the timeliness of the sensed information. The age of information (AoI) is a recently proposed measure of freshness or timeliness of the sensed information at the central controller. The problem becomes even more challenging under the denial of service attack from in-band full duplex jammers. Such jammers can sense the transmissions while injecting interference allowing them to react instantaneously with the transmission strategies. We develop a decision making policy for the central controller to combat such jammers with an aim to minimize the number of re-transmissions without compromising on the AoI. To boost the anti-jamming capability of the WSN, we introduce three modes for sensors: sleep, silent and transmit. We then jointly optimize the decision of when to switch the operating mode of each sensor and when to hop to a new channel using Markov Decision Process. We validate the superiority of the proposed strategy via extensive simulation results considering various parameters such as switching cost, transmission cost, failure to meet AoI cost and channel qualities.

The objective of this paper is to propose a computationally efficient load flow technique for an unbalanced active distribution network. A modified load flow technique is derived, which is generalized enough for considering various kind of distributed generation resource (DGR) with precise operating characteristics. Constant power voltage (PV) and droop regulated mode of operations are considered for DGR characteristics modelling. Proposed technique is suitable for weakly mesh feeder network since no separate treatment is required. Sequence component frame model is used in the network decoupling for faster computation. Tests are carried out on two different size network to verify accuracy and computation efficiency. Standard IEEE 34 and 123 bus feeder distribution network with some modifications are considered for the case study. Results are validated based on the computation speed and convergence achievement with literature.

Demand response (DR) technologies can curtail peak electricity usage, preserve energy conservation and, as well as reduce carbon emission by employing renewable energy sources. This paper presents appliances based home energy consumption scheduling in the framework of demand response technologies. The user behavior is accounted in the form of their daily energy need based on the appliance operation. The renewable energy sources such as solar photovoltaic and fuel cell are considered as generation available at the user homes and can be directed towards a significant reduction in carbon footprints. The energy user also sustains the use of electric vehicle and storage system. A day ahead load scheduling is done to allow the smart charging of electric vehicle and scheduling of smart appliances in order to reduce carbon emission and consumer daily bill benefits by indulging in the programs offered by the load-serving entity. The Simulation results demonstrate a significant reduction of 246.8 lbs in carbon footprints and maximum savings of$ 13.34 on users daily electricity bill when compared with before DR scheduling.

The objective of this paper is to improve the accuracy of the results obtained from the load llow analysis of a distribution network via forward-backward sweeps. Specific attention is paid to the two-port modeling of a transformer with precise consideration for the zero sequence components of its port voltages. The zero sequence voltages at transformer ports are often ignored in the conventional load llow analyses. A new two-port network model is derived, which is generalized enough for the accurate representation of a transformer in the cascaded connection. Based upon the novel two-port representation made, a new set of iteration rules is established to carry out the forward-backward sweeps for solving the load llow results. All possible transformer configurations are taken into account. It is shown that the load llow analysis technique proposed is suitable for both active and passive distribution networks. The accuracy analysis of the load llow results is also carried out. For a given load llow result, by assessing the nodal current imbalances are evaluated based upon the admittance matrix representation of the network. Case study is performed to demonstrate the utility of the proposed load llow analysis technique.