Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 3 of 3
Full-Text Articles in Controls and Control Theory
Drafting In Self-Timed Circuits, Christopher Cowan
Drafting In Self-Timed Circuits, Christopher Cowan
Electrical and Computer Engineering PhD Day
Intervals between data items propagating in self-timed circuits are controlled by handshake signals rather than by a clock. The sequence of handshakes can be abstracted as the movement of “tokens”. In many self-timed designs, a trailing token will catch up with a leading token, even when it trails by thousands of gate delays. Simulations in SPICE of a simple GasP circular FIFO reveal this effect. Contrary to earlier work, we find the cause of drafting to be charge stored on an isolated node between two series transistors. This mechanism occurs in many decision gates that implement a logical AND. The …
Coaxial Magnetic Gearbox With A Flux Concentration Halbach Rotor And Consequent Pole Rotor Typology, Ho Yin Wong
Coaxial Magnetic Gearbox With A Flux Concentration Halbach Rotor And Consequent Pole Rotor Typology, Ho Yin Wong
Electrical and Computer Engineering PhD Day
This paper presents the performance comparison between a flux concentration and triple permanent magnet consequent pole magnetic gear rotor typology with the same 7.5:1 gear ratio. The triple permanent magnet consequent pole magnetic gear is unique in that it can create torque change through two field heterodyning pole combinations. The performance with respect to volumetric and mass torque density is considered. It is shown that the flux concentration typology can operate with a higher mass and volumetric higher torque density, even when using a low flux concentration ratio rotor.
From A Locally Competitive Algorithm To Sensory Relevance Models, Walter Woods
From A Locally Competitive Algorithm To Sensory Relevance Models, Walter Woods
Electrical and Computer Engineering PhD Day
This poster addresses the development of a new Machine Learning (ML) mechanism, the Sensory Relevance Model (SRM), as a means of splitting information processing tasks into two sub-tasks with more intuitive properties. Specifically, SRMs are a front-end for other ML techniques, re-mapping the input data to a similar space with significantly more sparsity, achieved through the transformation and suppression of inputs irrelevant to the task. Prior work has attempted to reveal this information for Neural Networks (NNs) either as a post-processing step via saliency maps or through a simple masking of the input achieved with a dot product (so-called ``attention'' …