Analog-to-digital conversion with noise injection via wavefront multiplexing techniques

This application is a continuation of application Ser. No. 14/988,809, filed on Jan. 6, 2016, which is a continuation of application Ser. No. 14/637,405, filed on Mar. 4, 2015, now U.S. Pat. No. 9,246,508, which is a continuation of application Ser. No. 13/762,413, filed on Feb. 8, 2013, now U.S. Pat. No. 8,981,976, which is a continuation of application Ser. No. 12/985,044, filed on Jan. 5, 2011, now U.S. Pat. No. 8,384,572, which claims the benefit of U.S. provisional application Ser. No. 61/381,381, filed on Sep. 10, 2010.

1. Field

The present invention relates to architectures and designs of digital systems. More specifically, but without limitation thereto, the present invention pertains to an electronic signal conversion system that utilizes a noise injection system in order to maintain or increase signal resolution and increase the dynamic range. The present invention also offers a more time-efficient conversion as well as a more cost-effective conversion method.

2. Prior Art

The following is a tabulation of some prior art that presently appears relevant:

U.S. Patents Pat. No. Kind Code Issue Date Patentee 5,077,562 1991 Dec. 31 Chang et al. 5,630,221 1997 May 13 Birleson 6,049,251 2000 Apr. 11 Meyer 6,526,139 Bl 2003 Feb. 25 Rousell et al.

Non-Patent Literature Documents

Estrada, A.; Autotestcon, 2007 IEEE, “Improving high speed analog to digital converter dynamic range by noise injection”.

Currently in the electronics field, conversions between digital and analog signals are necessary for many day-to-day electronic operations. Analog signals are signals that utilize properties of the medium to convey the signal's information, essentially used in its original form. In particular for the field of electronics, an analog signal is taking a signal and translating it directly into electronic pulses. On the other hand, a signal is considered digital when it is processed into discrete time signals, usually in the form of a binary code (1s and 0s instead of a continuously variable function as found in analog signals). Nowadays, although nearly all information is encrypted digitally, analog signals commonly function as carrier signals for information transmission.

As a result, conversions between analog and digital signals for modern electronics are a common occurrence. For example, portable cellular phone signals are broadcast in the analog format and need to be converted to a digital signal within the phone itself for practical use.

Television signals are also transmitted in the analog spectrum and have to be converted to digital format for signal processing.

A key performance index of conversion from analog to digital (A/D) is the dynamic range, which is the ratio between the smallest and largest possible values of changeable quantities. Additionally, only signal strengths within the specified dynamic range can be detected. As a result, the dynamic range that is factored into A/D circuit design is required to be reasonably wide, and in some cases, to be as wide as possible. For instance, color perceptible to the human eye ranges from 4.28.times.10.sup.14 Hz (hertz) to 7.14.times.10.sup.14 Hz. If, for example, a TV's dynamic range cannot cover this spectrum, the quality of the TV signal will degrade as it cannot show all the colors in the received TV video signal.

Utilizing such wide dynamic ranges has several issues. While higher dynamic range means better precision and resolution of digital signals, the higher dynamic range also necessitates more expensive and precise equipment. There are cases where it is impossible to implement such devices either because it is impractical or too costly, such as in mobile devices.

Additionally, analog-to-digital conversions have an issue with unwanted noise being introduced into the signal. One source of noise is the conversion itself, as an analog signal is changed to a format that eliminates some of the fine resolution of the signal. Because of this, research has been performed to increase the dynamic range of analog-to-digital converters without changing the resolution, as well as reducing unwarranted and unwanted noise. The present embodiment of the invention aims to mitigate both of these factors in A/D converters by introducing a “noise” injection to essentially cancel out any unwanted noise as well as maintain a high dynamic range so that resolution is not lost in the conversion.

The present invention is a noise injection system for the purpose of eliminating unwanted noise while maintaining a high dynamic range for analog to digital conversions, comprising: a wave front de-multiplexer, multiple analog-to-digital converters and a wave front multiplexer.

The noise injection system performs as follows. Multiple input signal streams, noise injection streams, and a ground are all connected to a wave-front multiplexer, where the signal and noise signal outputs are connected to a multiplexer. Here, the signals are multiplexed (combined) into N data streams, each with a signal component of all inputs. The multiplexer output lines are transmitted to A/D converters. After conversion to digital format, the sampled digitized signals are transmitted to a wave-front de-multiplexer, where the data streams are recovered into output signals matching the inputs. These signals are then reconverted from digital to analog if necessary.

Through injecting noises which could be eliminated by filters afterwards, the present invention enhances signal strength while maintaining a high dynamic range. Weak signals out of the A/D converter dynamic range are now able to be detected because of added noise. In such a way, the signals' dynamic range is increased. Additionally, injecting noise also has the benefit of cancelling out any unwanted noise, thus increasing clarity and signal resolution.

An alternative embodiment of the present invention involves utilizing an optimization processor that is connected to the wave-front de-multiplexer. Samples of the signals being processed are sent to the processor, where an optimization loop adaptively adjusts the strength, phase, and wave front vectors of the noise in order to cancel out the unwanted noise. After processing, the signals are re-introduced into the signal streams for proper cancellation of unwanted noise.

With the proposed noise injection system, the dynamic range of the analog-to-digital conversion system can be accommodated with the injected noise level without redesigning the system. Furthermore, the signal converters in this invention process fewer bits of data, thus reducing power requirements, cost and complexity.

FIG. 1 is an illustration of an analog/digital conversion system with an attached optimizer

FIG. 2 is an illustration of an alternative implementation analog/digital conversion system

FIG. 3 is an illustration of another alternative implementation of the conversion system

The present invention relates to the architecture and design of electronic systems, and, in particular to electronic signal conversion hardware architecture and design.

An implementation of one embodiment is shown in FIG. 1 . In this particular embodiment, there are 4 input ports with 4 signal inputs including: incoming signal 102 a , injected noise signal input 104 a , and two grounded signals 105 a , 106 a , are connected to multiplexer 108 . The input ports in the actual implementation may vary, and not limited to 4 input ports. The injected noise signal 104 a and incoming signal 102 a will be split in wave-front multiplexer 108 and mixed with each other in order to improve dynamic range of the whole system. Ground 105 a and 106 a will be used as diagnostic signals.

Wave-front multiplexer 108 equally splits and mixes M input signals to form N output signals, where, in this embodiment, M and N are both 4 . Each of mixed N signals contains information from all M input signals. Each output of N signals maintains a fixed relative phase difference and N output signals form a wave front vector. For example, in case of FIG. 1 , if I use a 4-point Fast Fourier Transformer (FFT) as a wave front multiplexer, then the phase difference between each output signal is e −iπ/2 . The wave front vector is [1, e −iπ/2 , e −iπ , e −i3π/2 ]. This wave front vector will be used to recover the mixed signals.

Thus, after wave front multiplexer 108 processes the N inputs, 4 output signals are already incoming signals mixed with proper noises. If FFT is used as a wave front multiplexer, each channel only possesses A/D bandwidth of the original signal. As a result, cheap, low speed and low resolution A/D converters 110 a , 110 b , 110 c and 110 d are used to sample these signals. After conversion, the signals are all in the digital format.

A wave front de-multiplexer 112 performs the inverse process of wave front multiplexer. The de-multiplexer 112 is used to recover the mixed signals to the original input signals in the digital domain. For example, if FFT is used previously, an Inverse Fast Fourier Transformer (IFFT) will be used here. After this, an incoming signal in digital domain 102 b , an injected noise in digital domain 104 b , ground in digital domain 105 b and 106 b are recovered.

All signals are recovered due to the wave front vector which represents phase differences among signals. Therefore, if any distortion occurred in previous steps, the wave front vector will be distorted. However, with the help of optimizer 114 , even if signals are distorted, recovery can still be successful. By using diagnostic signals ground 105 a and 106 a , if signal recovery is successful, the recovered signals 105 b and 106 b should be perfectly zero. Optimizer 114 adaptively adjusts the wave front vector until the signals 105 b and 106 b reach zero. Thus, any previous distortion is compensated for, and the output signals exhibit improved clarity than without the present invention.

An alternative embodiment of the noise injection system is shown in FIG. 2 . Incoming signal 102 a and injected noise 104 a input signals in this embodiment. The rest of this embodiment is the same as the main embodiment. But optimizer, since there is no reference signal such as 105 a or 106 b , quality of the output signal cannot be determined.

Another alternative embodiment of the noise injection system is shown in FIG. 3 . The input signals include signal 102 a , injected noise 104 a and one grounded signal 105 a or 106 a . The rest of this embodiment is the same as main embodiment but optimizer. Signal 105 b can be used as a diagnostic signal. It is to indicate the quality of the output signal 102 b.