Noise canceller device for short-wave radioreceivers
Client’s Name (Country):
NDA
The Challenge:
To solve the problem of radio noise concentrated on the spectrum, various devices based on reject filter technology have been developed. However, using such filters to suppress narrow-band noise is economically impractical; this type of noise usually appears as a series of spectral lines and their cancellation would require an appropriate quantity of filters. Furthermore, reject filters demand preliminary adjustment due to the necessity of adding special block analysis attributes for a useful signal, or the presence of a live operator. Overall, there is a notable lack of effective devices with the task of suppressing arbitrary groups of narrow-band noise.
The Solution:
The AT Software team developed an entirely new signal processing algorithm, which differs from well known signal processing algorithms based on reject filters. The reasons for this are as follows: It does not require a priori information on current noise spectrum parameters; Its usage does not demand structural modification of the reception channel; The algorithm does not influence the operation of secondary signal processing devices. This new algorithm compensates noise concentrated on a spectrum n(t) at the reception of signals s(t) with amplitude modulation, frequency modulation or phase modulation on π. The proposed algorithm is based on a separate synchronous digitization of the received oscillations u(t) = s(t) + n(t) in null and max points of high-frequency useful signal infill, definition of the obtained indications spectrum, calculation of these spectrum modules and noise compensation with
subsequent restoration of the useful signal phase structure.
To check the new algorithm a prototype of the noise canceller based on a powerful DSP processor was developed. The algorithm was redesigned and moved from a float-point domain to a fixed-point domain, which effectively allows it to be put into a fixed-point DSP.
Tools and Technologies:
TMS 320C 6414 TI, Code Composer Studio
Project Duration (months):
11 months