Knowledge representation is key factor for problem representation particularly for incomplete information in expert systems. Usually incomplete information is fuzzy rather than likelihood. Learning methods are necessa...
详细信息
ISBN:
(纸本)9781509041114
Knowledge representation is key factor for problem representation particularly for incomplete information in expert systems. Usually incomplete information is fuzzy rather than likelihood. Learning methods are necessary to solve expert problems. The conditional inference method is studied different from Zadeh, Mamdani and TSK methods. Learning fuzzy conditional inference with individual methods fuzzy logic, Genetic algorithms, Petri net and neural net not sufficient for large problems. In this paper FuGePeNuNet method is studded by combining fuzzy logic, Genetic algorithms, Petri net and Neural net for large problems of Fuzzy Expert Systems. The fuzzy medical diagnose is given an example.
Newly-available spatial architectures to accelerate finite-automata processing have spurred research and development on novel automata-based applications. However, spatial automata processing architecture research is ...
详细信息
Newly-available spatial architectures to accelerate finite-automata processing have spurred research and development on novel automata-based applications. However, spatial automata processing architecture research is lacking, because of a lack of automata optimization and place-and-route tools. To solve this issue, we propose a new, open-source toolchain-Automata-to-Routing (ATR)-that enables design-space exploration of spatial automata architectures. ATR leverages existing open-source tools for both automata processing and FPGA architecture research. To demonstrate the usefulness of this new toolchain, we use it to analyze design choices of spatial automata processing architectures. We first show that ATR is capable of modeling the logic tiles of a commercially-available spatial automata processing architecture. We then use ATR to compare and contrast two different routing architecture methodologies-hierarchical and 2D-mesh-over a set of diverse automata benchmarks. We show that shallower 2D-mesh-style routing fabrics can route complex automata with equal channel width, while using up to 4.2x fewer logic tile resources.
Software testing is becoming so challenging work due to large-scale and high performance fieldprogrammable Gate Arrays (FPGAs) widely used in space utilization engineering projects. Though there are techniques such a...
详细信息
Software testing is becoming so challenging work due to large-scale and high performance fieldprogrammable Gate Arrays (FPGAs) widely used in space utilization engineering projects. Though there are techniques such as static analysis very popular in industry, no rounded testing framework or testing standard is constructed specifically for FPGA software. Based on traditional testing methods, this paper proposes a novel framework for FPGA testing which supports a through checking on the safety-critical applications. Test coverage is therefore improved by combining technologies including static timing analysis, behavior simulation and board-level testing. Experimental results on control software of Tiangong-2 Project demonstrate the proposed testing framework is valid yet efficient in high-speed signal testing and Triple Modular Redundancy (TMR) design validation.
Fuzzy logic based control system provides a simple and efficient method to control highly complex and imprecise situation. However, the lack of simple hardware designs that is the capable of implementing fuzzy control...
详细信息
Fuzzy logic based control system provides a simple and efficient method to control highly complex and imprecise situation. However, the lack of simple hardware designs that is the capable of implementing fuzzy controller's parameters in a single chip, sometimes limits the capability of fuzzy based system in an automotive and industrial environment. In present days, a reconfigurable hardware platform like fieldprogrammable gate array (FPGA) has become an alluring alternative to synthesis and implements the desired system with an appropriate design for automation tool. This paper aims to address different hardware implementation issues of fuzzy system on a real-time environment (FPGA) and also proposes the hardware friendly neuro-fuzzy technique as a better replacement of a fuzzy system. In this paper, a simple rule base has been realized on the Xilinx SPARTAN 3AN FPGA development board to bolster the feasibility and superiority of the neuro-fuzzy over the fuzzy system. For the purpose of hardware implementation, optimum choice in terms of speed, accuracy and resource utilization is obtained in case of neuro-fuzzy technique and the results are in conformation.
Machine-type communication (MTC) provides a potential playground for deploying machine-to-machine (M2M), IP-enabled 'things' and wireless sensor networks (WSNs) that support modern, added-value services and ap...
详细信息
ISBN:
(纸本)9781467389990
Machine-type communication (MTC) provides a potential playground for deploying machine-to-machine (M2M), IP-enabled 'things' and wireless sensor networks (WSNs) that support modern, added-value services and applications. 4G/5G technology can facilitate the connectivity and the coverage of the MTC entities and elements by providing M2M-enabled gateways and base stations for carrying traffic streams to/from the backbone network. For example, the latest releases of long-term evolution (LTE) such as LTE-Advanced (LTE-A) are being transformed to support the migration of M2M devices. MTC-oriented technical definitions and requirements are defined to support the emerging M2M proliferation. ETSI describes three types of MTC access methods, namely a) the direct access, b) the gateway access and c) the coordinator access. This work is focused on studying coverage aspects when a gateway access takes place. A deployment planar field is considered where a number of M2M devices are randomly deployed, e.g., a hospital where body sensor networks form a M2M infrastructure. An analytical framework is devised that computes the average number of connected M2M devices when a M2C gateway is randomly placed for supporting connectivity access to the M2M devices. The introduced analytical framework is verified by simulation and numerical results.
The possibility of improving on attributes of a solution which are traditionally opposed each other is proved. For digital components of safety-related systems, the method of improving on attributes of a checkability ...
详细信息
The possibility of improving on attributes of a solution which are traditionally opposed each other is proved. For digital components of safety-related systems, the method of improving on attributes of a checkability of the circuit and trustworthiness of the results calculated on FPGA with the LUT-oriented architecture is offered. The method is directed to improving of the ready project by a choice of the version of a program code without change of the hardware decision. Versions of LUT memory programming and a set of faults on which these versions exert impact are generated. Faults of shorts between adjacent address inputs of LUT are considered. Operation of the circuit is simulated on all versions and on all set of faults. The method selects the versions providing increase in a checkability of the circuit in a normal mode and trustworthiness of results in emergency mode of the safety-related systems.
We present a framework for distance-based classification of functional data. We consider the analysis of labeled spectral data and time series by means of Generalized Matrix Relevance Learning Vector Quantization (GML...
详细信息
ISBN:
(纸本)9781509066384
We present a framework for distance-based classification of functional data. We consider the analysis of labeled spectral data and time series by means of Generalized Matrix Relevance Learning Vector Quantization (GMLVQ) as an example. To take advantage of the functional nature a functional expansion of the input data is considered. Instead of using a predefined set of basis functions for the expansion a more flexible scheme of an adaptive functional basis is employed. GMLVQ is applied on the resulting functional parameters to solve the classification task. For comparison of the classification a GMLVQ system is also applied to the raw input data, as well as on data expanded by a different predefined functional basis. Computer experiments show that the methods offers potential to improve classification performance significantly. Furthermore the analysis of the adapted set of basis functions give further insights into the data structure and yields an option for a drastic reduction of dimensionality.
This paper presents enhancements to the Xilinx UltraScale+ clocking architecture to support fine-grain time-borrowing. Time borrowing improves performance by redistributing timing slack between fast and slow paths. Th...
详细信息
Heterogeneous computing is required in systems ranging from low-end embedded systems up to the high-end HPC systems to reach high-performance while keeping power consumption low. Having more and more accelerators and ...
详细信息
Convolution neural networks (CNNs) have become the state-of-the-art technique for Artificial Intelligence (AI) applications. But they consume resources heavily both in computation and energy. Compressing the parameter...
详细信息
Convolution neural networks (CNNs) have become the state-of-the-art technique for Artificial Intelligence (AI) applications. But they consume resources heavily both in computation and energy. Compressing the parameters to fewer bits is an effective way to reduce data quantity, making it possible to deploy CNNs on mobile, IoT and other resource-limited devices. The existing means need to re-train the entire networks and tune the parameters. This paper presents a convenient method to compress the parameters of trained CNN models without re-training and fine-tuning. Our method reduces the number of bits from 32 to 5-9 under the same networks and on the same dataset, with the loss of accuracy less than 1%. The logic power of a trained LeNet reduces by 70.58% on FPGA evaluation.
暂无评论