Digital Emulation of Pulse Frequency Modulation for Neuroprosthetic Sensory Feedback

Pulse frequency modulation (PFM) is a method of encoding information where the instantaneous frequency of a pulse train carries the signal's information. PFM is of particular interest to those working towards interfacing prosthetic devices directly with the human nervous system. In this paper, we consider the effects of directly implementing PFM with a digital microprocessor. We consider three digital PFM algorithms: two are deterministic, and the third has a probabilistic nature that has desirable time-averaged and ensemble behavior. For each algorithm, we analytically bound the error between the desired pulse frequency and the actual frequency output by the microprocessor. We aim to provide tools for the design and analysis of closed-loop neuroprosthetic systems containing PFM