Reconfigurable computing is being used to achieve high speed of Application Specific Integrated Circuits (ASICs), on the one hand, and the flexibility of the General Purpose Processors (GPPs), oil the other. However, ...
详细信息
ISBN:
(纸本)9781424415526
Reconfigurable computing is being used to achieve high speed of Application Specific Integrated Circuits (ASICs), on the one hand, and the flexibility of the General Purpose Processors (GPPs), oil the other. However, due to the requirement of multiple reconfigurations to complete a computation, the recontiguration overhead might degrade the performance of the system. In order to avoid excessive reconfiguration, this paper presents all Intelligent Reconfigurable Instruction Set Processor (IRISP). The proposed processor is based oil an Intelligent Computational Analyzer Unit (ICAU), which intelligently analyzes and compares the newly required configuration with that of the existing configuration of the processor and makes it possible to reuse maximum of the existing resources, while reconfiguring only the required resources. It has been shown that the intelligent reutilization of the existing resources significantly improves the performance of the processor.
Robotic systems are used in a wide range of applications. With an intuitive programming system from previous work we are able to transfer the advantages of robotic systems into more unstructured environments like priv...
详细信息
ISBN:
(纸本)9781538631577
Robotic systems are used in a wide range of applications. With an intuitive programming system from previous work we are able to transfer the advantages of robotic systems into more unstructured environments like private households or small and medium sized enterprises. An important challenge in this field is to enable non-experts to use all the capabilities of a robot. This includes two aspects: Robots must be intuitive to program and robust to execute. Our recently presented One-Shot programming framework concerns both aspects. The main contribution of this work is to extend the existing framework and to raise the usability of our approach. With this motivation we identify generalization chances in our task representation model and introduce an extension of our programming paradigm by integrating control flow to increase its capabilities. The proposed system concept enables users to guide a robot kinesthetically through a task without prior knowledge. By observing resources in the workspace, the demonstrated task is encoded as a finite state machine (FSM). This FSM allows the reproduction of a task by the robot itself and also offers two ways to control the flow of a robot program: Loops and Branches. With these statements a non-expert user gains the opportunity to program even more complex task solutions.
We present Versatile Inference Processor (VIP), a highly programmable architecture for machine learning inference. VIP consists of 128 lightweight processing engines employing a vector processing paradigm, with a simp...
详细信息
ISBN:
(纸本)9781728114446
We present Versatile Inference Processor (VIP), a highly programmable architecture for machine learning inference. VIP consists of 128 lightweight processing engines employing a vector processing paradigm, with a simple ISA and carefully chosen microarchitecture features. It is coupled with a modern, lightly customized, 3D-stacked memory system. Through detailed execution-driven simulations backed by RTL synthesis, we show that we can achieve online, real-time vision throughput (24 fps), at low power consumption, for both full-HD depth-from-stereo using belief propagation, and VGG-16 and VGG-19 deep neural networks (batch size of 1). Our RTL synthesis of a VIP processing engine in TSMC 28nm technology, using a commercial standard-cell library supplied by ARM, results in 18mm(2) of silicon area and 3.5W to 4.8W of power consumption for all 128 VIP processing engines combined.
This paper addresses the optimal kinematic control problem of the mobile manipulators. A computationally simple class of the Jacobian transpose control algorithms is proposed for the end-effector trajectory tracking. ...
详细信息
ISBN:
(纸本)9781538639269
This paper addresses the optimal kinematic control problem of the mobile manipulators. A computationally simple class of the Jacobian transpose control algorithms is proposed for the end-effector trajectory tracking. These controllers use a new non-singular Terminal Sliding Mode (TSM) manifold, as being a non-linear integral mapping of the second order with respect to the task space tracking error. Based on the Lyapunov stability theory, Jacobian transpose control schemes proposed are shown to be finite-time stable provided that some reasonable assumptions are fulfilled during the mobile manipulator movement. The performance of the proposed control strategies is illustrated through computer simulations for a planar mobile manipulator which accomplishes trajectory tracking by the end-effector in a two-dimensional task space and simultaneously minimises the Euclidean instantaneous distance between initial configuration and the current one.
As programmable logic is conquering data centers, FPGA resources become a ubiquitously available commodity available through the cloud. In these ecosystems, their employment must advance to fit the role of a dynamical...
详细信息
ISBN:
(纸本)9781538655221
As programmable logic is conquering data centers, FPGA resources become a ubiquitously available commodity available through the cloud. In these ecosystems, their employment must advance to fit the role of a dynamically managed, shared resource within multi-user environments. programmable hardware vendors provide engineering workflows that only start to acknowledge the requirements of this thriving application context, which makes resource regularization, design migratability and isolated execution imperatives. Quite a few academic efforts have aimed at the key technique of design relocation on top of the vendor-provided workflows. Regularly, critical shortcomings remain and, in all cases, the engineering procedure is ridiculously complex. This paper identifies the significant gaps left by previous approaches and describes a thoroughly automated workflow for building a system infrastructure on FPGA that makes it a resource that is manageable by the OS. It highlights an end-to-end workflow to enable (a) the out-of-context synthesis of end-user designs against a thin module interface defined by appropriate synthesis constraints, which are (b) universally employable to all the user slots provisioned on an appropriately partitioned device without re-synthesis and (c) will operate there appropriately isolated merely communicating to the user host application through privately allocated communication channels.
A novel approach for efficient implementation of applications onto reconfigurable architectures is introduced. The proposed methodology can applied both for designing an interconnection architecture as well as for mak...
详细信息
ISBN:
(纸本)3540390944
A novel approach for efficient implementation of applications onto reconfigurable architectures is introduced. The proposed methodology can applied both for designing an interconnection architecture as well as for making a thermal-aware placement. In the first case, the dominant parameters that affect performance and energy (segment length and switch boxes) are examined. This approach is based on finding the optimal wire length and then making exploration in order to determine the appropriate combination of multiple switch boxes. In the second case, a new technique for thermal-aware placement is introduced. The main goal of this technique is to spread out the power consumption across the whole device, as well as to minimize it. Both of the methodologies are fully-supported by the software tool called EX-VPR. For the purposes of this paper, the EnergyxDelay Product (EDP) is chosen as selection criterion for the optimal interconnection network, while the total power consumption is the criterion for the thermal-aware routing. For the designing of the interconnection network we achieved EDP reduction by 45%, performance increase by 40% and reduction in total energy consumption by 8%, at the expense of increase of channel width by 20%. On the other hand, for the thermal-aware approach, we spread the heat and power across the whole FPGA, while we achieve about 20% reduction in total power consumption. In this case, the penalty in channel width is about 10%.
This article considers the robust cooperative control problem of multiple double-integrator agents' formation moving around a set of given curves on spheres, where each agent suffers an unknown spatiotemporal flow...
详细信息
ISBN:
(纸本)9781538626795
This article considers the robust cooperative control problem of multiple double-integrator agents' formation moving around a set of given curves on spheres, where each agent suffers an unknown spatiotemporal flowfield and the communication topology among agents is directed. The flow specification is composed of a general direction matrix and a responding unknown flow speed vector, which includes almost all forms of flowfield in the literature. A new second-order adaptive update law is given for estimating the flow speed vector by using the tool of adaptive backstepping, which is incorporated into our previous geometric extension design for the purpose of designing the robust spherical formation tracking control law. A potential function is used to avoid each agent's movement to the undefined angle on the points between the poles of sphere. The asymptotic stability of the system is proved when the directed communication topology is strongly connected. The effectiveness of the analytical results is verified by a numerical simulation.
Realistic reference networks are critical for all techno-economic evaluations in the design of an optical network architecture. As this information was not publicly available for Ireland, we generated such a reference...
详细信息
ISBN:
(纸本)9781538608593
Realistic reference networks are critical for all techno-economic evaluations in the design of an optical network architecture. As this information was not publicly available for Ireland, we generated such a reference network. In this paper we elaborate on the challenges faced during the generation and present some statistics. We also show how this is used in the design of an optical architecture for the same country. The level of detail of the reference network needed varies according to the layer of the architecture that is being designed. In this paper we show how the original reference network is reduced while maintaining important properties for the design of the architecture. We also elicit geotypes from the structure of the reference network and discus the importance of this computation in the design of the optical distribution network.
In this paper, we consider the problem of scheduling an agile sensor to perform an optimal control action in the case of the multi-target tracking scenario. Our purpose is to present a random finite set (RFS) approach...
详细信息
ISBN:
(纸本)9780996452700
In this paper, we consider the problem of scheduling an agile sensor to perform an optimal control action in the case of the multi-target tracking scenario. Our purpose is to present a random finite set (RFS) approach to the multi-target sensor management problem formulated in the Partially Observed Markov Decision Process (POMDP) framework. The reward function associated with each sensor control (action) is computed via the Expected Risk Reduction between the multitarget predicted and updated densities. The proposed algorithm is implemented via the Probability Hypothesis Density filter (PHD). Numerical studies demonstrate the performance of this particular approach to a radar beam-pointing problem where targets need to be prioritized.
Recently (authors, POPL 2017), a notion of dimensionality for intersection types was introduced, and it was shown that the bounded-dimensional inhabitation problem is decidable under a non-idempotent interpretation of...
详细信息
ISBN:
(纸本)9781509030187
Recently (authors, POPL 2017), a notion of dimensionality for intersection types was introduced, and it was shown that the bounded-dimensional inhabitation problem is decidable under a non-idempotent interpretation of intersection and undecidable in the standard set-theoretic model. In this paper we study the typability problem for bounded-dimensional intersection types and prove that the problem is decidable in both models. We establish a number of bounding principles depending on dimension. In particular, it is shown that dimensional bound on derivations gives rise to a bounded width property on types, which is related to a generalized subformula property for typings of arbitrary terms. Using the bounded width property we can construct a nondeterministic transformation of the typability problem to unification, and we prove that typability in the settheoretic model is PSPACE-complete, whereas it is in NP in the multiset model.
暂无评论