![]() |
Adaptive Integrated Microsystems Laboratory |
|
"Towards Hybrid, Smart Circuits and Systems" Our research group focuses on three aspects of hybrid circuits and systems: MORPHING - Investigating neural inspired circuits; SYNTHESIS - Using hybrid computational elements (biological and silicon) to design biomolecular circuits and systems; MONITORING - Embedded and implantable monitoring of natural and engineered systems. |
| Members Research Publications Patents Courses Download |
|
|
On-chip Learning: Smart ADC (more...) |
Open-source Mixed-signal Test-station Platform A key component in mixed-signal design is a reliable test-station, which can be used for evaluating and benchmarking the performance of fabricated chips. Cost and time is a critical factor, especially for researchers who are in the process of establishing programs in mixed-signal testing. Also, obtaining measurements in field the test-station has to be portable and has to be scalable to support most existing interfaces. We have designed a low-cost (estimated cost less than $1000) and portable test environment which is easy to use and can interface with popular data analysis and testing software like MATLAB and LabView. The design and software files for the test-station are available for free, with the hope that you will expand the functionality of the station, so that other researchers could benefit from it.
Gini-SVM
is a multi-class classifier that directly produces conditional
probability estimates and hence can be used for data-fusion and for
integrating with other probabilistic models like Hidden Markov Models
or Gaussian Mixture Models. The code is available for download here (MATLAB version) and (C version will be available soon). The Gini-SVM JMLR paper can be downloaded (here).
Speech Processing Software Sigma-delta
learning scripts for speech processing and super-resolution source
separation can be downloaded (here). The sigma-delta learning algorithm
is described (here) and the super-resolution source separation is described in the paper accepted in IEEE Transactions of Signal Processing.
![]() FAST: Factor-graph Analysis of Stochastic CircuiTs. FAST
is a Monte-Carlo analysis tools for estimating the reliability of
circuits whose components are stochastic elements for example protiens
and Biomolecules. The documentaion describing the FAST interface is
provided in the following (slides). The TBioCAS paper describing the biomolecular analysis using factor-graphs can be downloaded (here).
MarginLDPC decoders The
C code comparing the performance of three different decoding algorithms
used for Low-density Parity Check can be downloaded (here). The paper
describing the results of the comparison for a representative LDPC code
can be downloaded (here).
Other Useful Tools Power-point circuit drawing Tool (Download Link)
|
|||||||||
|
|
Energy harvesting Circuits and systems (more..) | ||||||||||
|
|
Energy efficient Analog LDPC decoders (more...) |
||||||||||
![]() |
Forward Error Correcting Biosensors (more...) | ||||||||||
|
|
Acoustic Sensor Arrays (more...) | ||||||||||
![]() |
Multi-channel data acquisition and potentiostats (more...) | ||||||||||
|
|
Learning and signal processing algorithms (more...) | ||||||||||
|
Questions and comments: shantanu at msu dot edu |