The autonomous self-healing (eDNA) hardware platform is a reconfigurable field-programmable gate-array (FPGA)-type platform developed by Technical University of Denmark (patent: WO/2010/060923). It is capable of autonomously reconfiguring itself in case a fault is detected and, thusly, restoring functionality at a fault-free location on the chip.
The software implemented on the liquid crystal waveguide Fourier transform spectrometer (LCW-FTS) prototype has been ported to the eDNA platform, resulting in an LCW-FTS processing system that can repair itself in case a fault is injected. The FPGA-based eDNA prototype has been ported to the FPGA of the embedded system of the FTS.
As transistor geometries continue to shrink, the variability of them increases. This results in an increasing number of both permanent and transient hardware faults, which in turn increases the demand for robust hardware systems. Particularly, the space environment stresses the hardware onboard a spacecraft. eDNA architecture has three key components: the electronic DNA (eDNA), an array of electronic cells (eCells) connected in a 2D-mesh network-on-a-chip (NoC), and a subset of eCells in this array that are designated “spare eCells.”
Analogous to the way biological cells work, all eCells interpret the eDNA and, based on their position, implement a part of the eDNA. Spare eCells are eCells that determine they should not implement any part of the application. The main benefit of the eDNA architecture is its capability to self-heal. The eCells use a cooperative self-test algorithm to ensure detection of permanent and transient faults. Once a fault is detected, self-healing is initialized. Each eCell keeps in its memory a list of spare eCells. When an eCell detects a fault, it determines the closest spare eCell by calculating the Manhattan distance, and then notifies the eCell that this spare eCell will be used to overtake the functionality of the faulty eCell. Then, it sends a package to the spare eCell, which assigns the cell number of the faulty eCell to the spare. Upon reception of the package, the spare eCell resets its cell number and runs the self-organization algorithm and takes on the functionality of the faulty eCell.
At the time of this reporting, this is the first time control and data processing of an FTS instrument has been implemented in an autonomous, self-healing hardware platform. Self-healing electronics represent a tremendous advantage in space applications where repair is very expensive, impossible, a high risk, or a combination of the three.
This work was done by Thomas T. Lu, Didier Keymeulen, and Tien-Hsin Chao of Caltech; and Jan Madsen and Michael R. Boesen of Technical University of Denmark for NASA’s Jet Propulsion Laboratory. NPO-47896
This Brief includes a Technical Support Package (TSP).
Fourier Transform Spectrometer on Autonomous Self-Healing Hardware Platform
(reference NPO47896) is currently available for download from the TSP library.
Don't have an account? Sign up here.