Multiplier for DSP Application in CPS System

Multiplier for DSP Application in CPS System

Abhishek Kumar
Copyright: © 2021 |Pages: 22
DOI: 10.4018/978-1-7998-5101-1.ch011
(Individual Chapters)
No Current Special Offers


A cyber-physical system over field-programmable gate array with optimized artificial intelligence algorithm is beneficial for society. Multiply and accumulate (MAC) unit is an integral part of a DSP processor. This chapter is focused on improving its performance parameters MAC based on column bypass multiplier. It highlights DSP's design for intelligent applications and the architectural setup of the broadly useful neuro-PC, based on the economically available DSP artificial intelligence engine (AI-engine). Adaptive hold logic in the multipliers section determines whether another clock cycle is required to finish multiplication. Adjustment in algorithm reduced the aging impact over cell result in the processor last longer and has increased its life cycle.
Chapter Preview


Artificial Neural Networks (ANN) depend on the simultaneous design and inclined by the human mind. ANN is a type of multi-processor program involving essential handling components known as neurons; The ANN has a complex network and versatile association between underlying components. The very first usage of ANN accompanied by the creation of the perceptron set out and related learning rule by Frank Rosen Blatt (Martin, 2002). Another critical advancement was incorporating the algorithm like back-propagation for training (Ajith Abraham, 2005). Image compression includes decreasing the measure of the memory expected to store a digital format. Aside from the current innovation for picture pressure, for example, JPEG, MPEG, and H.26x gauges, recent changes of neural systems are essence investigated. Fruitful utilizations of neural networks to vector quantization gotten settled. Different parts of neural system contributions around there are venturing up to assume significant jobs in helping with conventional compression methods (Venkata Rama Prasad Vaddella and Kurupati Rama, 2010).

The cyber-physical system (CPS) presented by (Ding 2019 and Kinsey, 2011); is a computational system organized of computational devices and physical conditions. Computational methods communicate through a communication network and control physical situations through actuators and can receive feedback input about physical conditions employing sensors. The collaborator of a CPS should know about every device that impacts the functionality of the CPS.CPS framework models usually comprise of a large number of differential conditions. The synthesis of such complex equations using software on a programmable chip is moderate. A few past efforts to implement; models as equivalent circuits on Field-Programmable Gate Arrays (FPGAs), showing a massive increase in speed, because of the magnificent tie among fine-grained local communication, which is the prime focus in physical models and the fine-grained parallel computation component and connection network of FPGAs. A CPS system includes security in terms of software protocol. Hardware implementation of complex computation comes with the limitation of leakage issue, cyber-attack says during computational device leaks side information. Intruder utilizes this information to get access to the device. A conventional cybersecurity mechanism cannot detect or block such a category of attack (Greenword, 2005). Recently, Artificial Intelligent (AI) is a work in progress to protect the CPSs.

Concerning digital security, AI advancements are utilized to measure information that originates from various sources of data. Artificial neural system is a framework that accepts the input, processes according to an algorithm, and furnishes results. Two primary aspects of machine learning are preparing and deduction of the data. Deduce with numerous varieties of typically lower exactness multipliers, gives off an impression of being a decent counterpart for FPGA models. Current FPGAs (Langhammer, M., & Baeckler, G., 2018). now have over a million lookup tables (LUT) mixes, which proposes that they are appropriate for executing these sorts of exhibits. Training requires, requires higher precision arithmetic computations.

The chapter is organized as follows; background highlights the needs of the DSP multiplier for the Xilinx AI engine, the MAC unit with aging indicator provided in the implementation of MAC, and the application of MAC with different types Verilog HDL. Simulation result with the conclusion and future direction has been discussed in the consecutive section.

Key Terms in this Chapter

DSP System: Is a dedicated processor to perform a wide variety of signal processing operations.

Ageing: Effect of the hardware block represents the degradation into parameter in % concerning time. HCI and BTI effect on underlying transistors estimate the fall of reliability of the complete circuit.

MAC: Majorly used block in DSP system, it computes the product of two number and adds that product to an accumulator

FPGA: A semiconductor device prototypes the digital system. It is based around with configurable logic block connected via programmable interconnect. FPGA is reconfigurable.

Compressors: Used in this chapter implement the large size of the multiplier. It maps maximum input to minimum output; the primary cell of the compressors is an adder.

Synthesis: Is a process of transforming HDL design into a gate-level netlist.

AI Engine: Xilinx’s AI engine provides higher computation density for neural network algorithm implementation. These AI engines include SIMD and VLIW dedicated RISC scaler processor. It can perform 512 b fixed and floating-point computation.

Multipliers: Is one of the key hardware blocks in most digital signal processing (DSP) systems. Power dissipation and delay are the primary design constraint which can be optimized by enhancing the performance of underlying multipliers.

Complete Chapter List

Search this Book: