The channels over which data is transmitted can cause signal distortions and subsequent noise. Often before the data is transmitted, it must first be processed by a time encoding machine, which traditionally can only handle single types of signal modalities. As a result, sending signals of different dimensions, i.e. audio and video, requires multiple encoders and multiple channels, thus resulting in increased inaccuracies. This technology describes a system for encoding multiple signals of different dimensions together into a single signal. Once the signal has been transmitted, this technology then uses channel identification upon decoding to filter out distortions that have been introduced by the transmission channel. It is compatible with a wide range of circuit types, from communications systems to neural circuits, thus providing improved reliability and capabilities for signal transmission.
This technology enables both asynchronous encoding and decoding of input signals that have different modalities and dimensions. Allowing for it to be integrated with various types of signals and applications, it can be realized by using a number of different asynchronous encoders, such as analog to digital converters, oscillators with additive or multiplicative coupling, or threshold and fire neurons. Channel identification is performed by splitting input signals, and then reconstruction coefficients are determined and used for signal reconstruction.
Numerous different implementations of this technology have been described, involving audio & video integration, multisensory integration, and IAF neurons.
Patent Pending (WO/2014/190155)
Tech Ventures Reference: IR CU13312, CU13313, CU13318