In 1998, scientists studying telemetry from two historically significant robotic spacecraft (Pioneer 10 and Pioneer 11) noticed something odd:  both craft were traveling more slowly than expected.  They were experiencing a steady deceleration of about a billionth of a meter per second squared – vanishingly small from our regular earthbound perspective, but significant in the ultra-precise high-speed long-duration world of interplanetary navigation.  Even when mission planners took such tiny factors as the gravitational attraction of the Sun and each planet, radiation pressure from the Sun, tiny thrust from possible fuel leaks, and the Solar Wind into account, they still could not account for this discrepancy which came to be known as the Pioneer Anomaly.  What added to the puzzle (and held the key to the answer) was that none of the other craft launched before or after showed this effect.  They all behaved exactly as predicted.  But now a team led by Frederico Francisco from the Institute for Plasmas and Nuclear Fusion in Lisbon reckon they may have solved the puzzle.

Over the years, a great many ideas have been suggested, from the mundane to the exotic.  Some like the suggestion that there was a flaw Einsteins theories on gravitation, which remains our best fit, seemed to be grasping at straws since they ignored the fact that the anomaly was unique to these two spacecraft.  Other scientists suggested that the measurements themselves could be suspect.  One idea, which was rejected early on, was that infra-red radiation from the craft’s own internal heat (generated  mainly by the radio-isotope power supply) could exert a tiny amount of radiation pressure, which would cause a tiny amount of thrust of about the same magnitude as observed.
What Francisco’s team did was to revisit this last idea, using a novel approach.  In order to properly model the emissions of infra-red light, they used a technique from the world of computer graphics:  Phong Shading.  Developed by Dr Bui Tuong Phong in 1970, this technique has become a standard in graphical rendering software to accurately model light reflecting from curved surfaces so as to draw realistic shading.  When Phong Shading was applied to 3D computer representations of Pioneer 10 and Pioneer 11, Francisco got a very accurate model of the infra-red emissions.  With this new data they were able to show that the radiation pressure from the emitted heat, plus the pressure from that same heat reflecting from large surfaces (like the antenna dish) could accurately account for the anomalous acceleration. 
An impressively novel way to use unusual tools to solve an old problem!  The paper describing their work can be downloaded and read here: 
http://arxiv.org/pdf/1103.5222v1