In the realm of neuroprosthetics, the imperative for hardware-efficient, faster spiking neuron models, reproducing spiking patterns with better precision, has emerged as a pivotal necessity; catalyzing advancements towards more compact and resource-optimal solutions for seamless integration within neural prosthetic systems. In this brief, a new digital model for Izhikevich Spiking neuron is proposed. The model accurately replicates the spiking patterns and phase portraits corresponding to the neuronal dynamic behaviour with a very high matching between the original and proposed models, high speed and improved hardware efficiency. The results of synthesis on the field programmable gate array (FPGA) platform demonstrate that the proposed hardware gives significant improvements in resource consumption and maximum operating frequency when compared to the latest existing Izhikevich spiking neuron architectures. A digit classification system using the proposed SNN model is implemented to demonstrate its applicability. Polychronous neuronal groups, which are ideal for signal-processing tasks like language learning and object recognition, are also simulated using the proposed model, which is significant in neuroprosthetic applications.

This article proposes a new activation function (AF)$\mathrm {sinh}(\beta x) + \mathrm {sinh}^{-1}(\beta x)$called combined hyperbolic Sine (comb-H-sine) to replace existing AFS like$\sigma $and$\tanh $in long short-term memory (LSTM) neural networks. The comb-H-sine function is implemented using purely combinational architectures with a 16-bit data width with fixed-point representation, resulting in improved accuracy. Both software and hardware modeling are used to investigate the proposed architecture. Compared to prior works on$\sigma $and$\tanh $functions, the hardware for the comb-H-sine function shows significant improvements in power consumption, number of cells, cell area, and delay. The proposed LSTM architecture using comb-H-sine outperforms existing AFs in terms of power delay product and accuracy on various datasets.

In this paper, a new multiplier-free transformation is proposed for the design of low complexity two dimensional (2D) finite impulse response (FIR) filters with improved circularity at wideband radii. By designing the 1D filter multiplier-free and using proposed multiplier-free transformation, a 2D circularly symmetric multiplier-free filter can be realized. Accuracy of the circular contour is measured based on contour approximation error. To prove the hardware efficiency of the proposed transformation, a 2D band-pass filter and a low-pass filter are designed and the overall complexity in terms of the number of multipliers required to implement the 2D filters is compared. From the results it is clear that the proposed transformation have good circularity at wideband radii and have less implementation complexity.

A computationally efficient maximally decimated filter bank core, suitable for digital hearing aids is designed and implemented in this work. The main advantages of the proposed design method are that the computational complexity and delay are less than that of the other filter bank based architectures in the literature. To reduce the hardware complexity of the filter bank core, filter coefficients are optimized and are implemented multiplier-free.