Automatic Fault Handling on Coarse Grain Reconfigurable
Computing Architecture
Liguo Song1,2
1. Department of Automatic Control, Beijing University of Aeronautics and Astronautics,
Beijing, 100083, China
2. Beijing Institute of Micro-electronics Technology, Beijing, 100076, China
songlg123456@sohu.com
Abstract: Conventional circuit design exploits device characteristics with a
certain temperature/radiation range, when this is exceed, the circuit may not
damaged. For Evolvable Hardware (EHW), although one or more elements
maybe damaged, the function can be recovered by utilizing residual
resource. An additional DSP (digital signal processor) is usually used as a
controller, which runs some algorithm for a long time to find a way to
handle fault. This paper describes a new evolvable coarse grain
reconfigurable platform currently under development. This platform differs
significantly from previous efforts, in that it has a coarse-grained,
signal-flow architecture oriented towards implementing multi-media and
stream-media. If some fault occurs, it can quickly find out which element is
error and automatically running recovery program to comeback function.
Key Words: evolvable hardware, coarse grain, reconfigurable computing
1 Introduction
Future space mission to Moon, Mars and beyond will face extreme environments
(EE), including environments with large temperature swings and high radiation levels.
EE can induce damage into electronic devices. The current approach for space
electronics designs is to use military range electronics protected through passive or
active thermal control, and high weight shielding for radiation reduction. This adds to
sizable weight and volume, compounded by power loss and additional cost for the
mission.
A new approach to alleviate the damage on electronic devices in EE is by
applying re-configurable devices and adaptive self-reconfiguration of circuit topology.
This approach refer here as although some parts or modul