June 6, 2016
Behind every successful spacecraft mission is a team of experts running thousands of calculations to accurately predict what could go wrong and avoid costly mistakes.
For this level of mathematical prediction, NASA is turning to Brandon Jones, an uncertainty quantification researcher and aerospace engineering professor in the Cockrell School of Engineering. Jones has developed a suite of uncertainty quantification tools outfitted with special algorithms that can run thousands of calculations to deliver predictions or simulations. The major advantage: Jones’ tools run on a regular desktop computer. NASA is now using his tools so that its spacecraft can avoid colliding with other objects and debris.
Engineers and scientists rely on uncertainty quantification — the science of characterizing and reducing uncertainties by determining the likelihood of certain outcomes — for everything from crash-testing cars to launching rockets.
Other commonly employed uncertainty quantification tools require the computational power of a super computer because they rely on the Monte Carlo method, a technique in which a large quantity of randomly generated and propagated samples are studied using a probabilistic model to find an approximate solution. The Monte Carlo method can be time consuming and costly.
“We think our tools have the potential to be faster and just as accurate as existing tools that require a super computer,” Jones said. “Our tools could potentially be used in diverse industries, including the automotive industry, where autonomously making decisions under uncertainty will become critical to integrating driverless cars into our transportation systems.”
The convenience of running predictive models on a desktop computer has already attracted the attention of the nation’s space organizations.
NASA engineers are using Jones’ tools on a daily basis for the ongoing Magnetospheric Multiscale, or MMS, Mission, to ensure that the collection of four spacecraft, which are flying in a precise formation, won’t collide when traveling near Earth.
And the NASA Jet Propulsion Laboratory (JPL) is working with Jones and his tools to potentially help design the Cassini Solstice spacecraft’s fly-bys of Titan, Saturn’s largest moon. In a typical fly-by, several positioning maneuvers are necessary to place the spacecraft on the right flight path and to account for any errors.
If JPL positions the Cassini Solstice correctly, it can leverage the moon’s gravitational field like a sling shot to turn the spacecraft onto a desired path.
“We are working on the tools that account for all these uncertainties, so when you design your maneuver to do your fly-by, you design the maneuver such that you minimize or control your error,” Jones said. “That requires you to account for all the uncertainties in the problem.”
Ultimately, Jones’ goal is to create a suite of tools that can be folded into existing systems. The research is funded by NASA’s Goddard Space Flight Center and JPL, as well as the Air Force Research Laboratories.
