University of Arizona researchers are developing hybrid hardware/software systems that could eventually use machine intelligence to allow spacecraft to fix themselves. Arizona professor Ali Akoglu is using field programmable gate arrays (FPGA) to build self-healing systems that can be reconfigured as needed to emulate different types of hardware. Akoglu says general-purpose computers can run a variety of systems but they are extremely slow compared to hard-wired systems designed to perform specific tasks. What is needed, Akoglu says, are systems that combine the speed of hard-wired systems with the flexibility of general-purpose computers, which is what he is trying to accomplish using FPGAs. The researchers are testing five wirelessly-linked hardware units that could represent a combination of five landers and rovers on Mars. Akoglu says the system tries to recover from a malfunction in two ways. First, the unit tries to fix itself at the node level by reprogramming malfunctioning circuits. If that fails, the unit tries to recover by employing redundant circuitry. If the unit's onboard resources cannot fix the problem, the network-level intelligence is alerted and another unit takes over functions that were done by the broken unit. If two units go down, the three remaining units divide the tasks. "Our objective is to go beyond predicting a fault to using a self-healing system to fix the predicted fault before it occurs," he says.