Advanced Signal Processing for Communication Networks and Industrial IoT Machines Using Low-Cost Fixed-Point Digital Signal Processor

Abstract

In this paper, the uses and functions of an existing, widely available digital signal processor (DSP) is extended to include using it to generate needed waveforms that could be used for computing, filtering, modulation and noise removal applications at the edges of communication networks that supports Industrial Internet of Things (IIoT) and Cyber Physical Systems (CPS). Such waveforms could also be used to optimize performances of IIoT communication networks such as wireless and powerline communication networks. The C28x digital signal processor (DSP), manufactured by Texas Instruments is used as a case study, and it is deployed in this paper as a mother wavelet generator by programming it to generate needed wavelets using embedded C programming language. Our implementation is the first known application of the C28x as a basic mother wavelet generator. Advanced signal processing method which include signal clipping, sequencing and concatenation are used with embedded C to program the C28x DSP. Wavelets generated with the C28x DSP are found to satisfy the wavelet admissibility condition. The open-loop voltage signal of an IIoT machine, generated at an IIoT network edge, and sent across a powerline communication (PLC) channel is used to evaluate the performance of the constructed Mexican Hat wavelet. The Mexican Hat wavelet constructed using the C28x DSP is used to remove noise from the machine signal corrupted with noise when the signal is transmitted over a PLC channel. Denoised signal using the DSP based Mexican Hat wavelet shows high correlation when compared to the original transmitted signal, while communication channel noise is successfully removed from the original signal.