Abstract: Systems and methods increase power efficiency in communication systems by examining a digital signal to determine whether a threshold corresponding to an increase in a power requirement is likely to be exceeded. The signal is encoded with information indicating the likely change and communicated to a driver that, upon extracting the information, uses it to cause instruct an amplifier to increase a power output to accommodate the increase in power requirement. Once the threshold is no longer exceeded, the driver circuit, advantageously, decreases the power output to conserve energy. In various embodiments, an amplifier may increase power efficiency by switching from a low-power circuit configuration to a high-circuit configuration.

Abstract: Systems and methods increase power efficiency in communication systems by examining a digital signal to determine whether a threshold corresponding to an increase in a power requirement is likely to be exceeded. The signal is encoded with information indicating the likely change and communicated to a driver that, upon extracting the information, uses it to cause instruct an amplifier to increase a power output to accommodate the increase in power requirement. Once the threshold is no longer exceeded, the driver circuit, advantageously, decreases the power output to conserve energy. In various embodiments, an amplifier may increase power efficiency by switching from a low-power circuit configuration to a high-circuit configuration.

Abstract: Storage-efficient, low-cost systems and methods provide embedded systems with the ability to dynamically perform on-device learning to modify or customize a trained model to improve computing and detection accuracy in small-scale devices. In certain embodiments, this is accomplished by repurposing storage elements from inference to training and performing partial back-propagation in embedded devices in the final layers of an existing network. In various embodiments replacing weights in final layers, while using hardware components to iteratively performing forward-propagation calculation, advantageously, reduces the need to store intermediate results, thus, allowing for on-device training without significantly increasing hardware requirements or requiring excessive computational memory resources when compared to conventional machine learning methods.

Abstract: A synchronized sampled data system comprising two or more sampled data circuits (10) each sharing a common databus (12) providing instructional words. Each sampled data system has an output (20) synchronized and phase shifted with respect to the other by preloading a unique start word into its respective internal counter (22) via a common control line. After initialization, each counter is clocked in synchronization with the other by a common clock (SCLK), such that the counter's counts remain the predetermined count difference from the other, and respective circuit output (20) maintains the corresponding phase relationship. The present invention finds particular advantages controlling switching power supplies and A to D converters. A desired phase lag can be established with a granularity of 1/M clock cycles. The present invention achieves lower input ripple current, lower in-band conducted emissions and better signal to noise ratio of the whole system.