SERVICES & SYSTEMS INTEGRATION

EO/IR

Summary

Challenge: 

  • Port lab-developed code to embedded processors
  • Interface to multiple sensors
  • Meet strict timing requirements

Result: 

  • Met all technical requirements
  • Completed flight trials

What we did: 

  • Map the application
  • Develop a novel chassis
  • Develop a new customized module (IOM212)
  • Build sensor interface IP, integrated IP from ourselves and two other partners
  • Built the full image processing application
  • Remapped the algorithms to the hardware/optimized them, etc.
  • Brought unit through lab qual
  • Supported flight qualification

Full Story

The customer had a complex set of algorithms that needed to be quickly deployed on an airborne, POD-based platform. The problem was, the existing application literally required racks of sensor-processing hardware to manage the heavy processing requirements and did not operate in real-time. Compounding this issue, a QRC requirement demanded quick deployment – while delivering on critical battlefield needs.

Mercury was selected to provide not only the real-time image processing system, including the image processing application, but also the computer to manage sensor/processor tasking, and the onboard storage system.  Additionally, the Mercury team, which had to perform their own development, also performed integration activities with both the prime and three other team members. This further challenged the team by having to manage five simultaneous development efforts. As is common with such efforts, Mercury faced a multi-dimensional problem set including balancing overall SWaP constraints with performance levels and overall functionality along with overall schedule risk. We responded with a unique solution that was smaller and more efficient than systems that had been previously developed – by one estimate, the system that Mercury developed was supposed to be “impossible,” as it processed 15x more pixels in the same SWaP with the same real-time constraints than the competing solution.

Additionally, Mercury was responsible for mapping the application to the new hardware, developing a novel chassis, developing a new customized module (IOM212), building sensor interface IP, integrating IP from ourselves and two other partners and building the full image processing application. We then remapped the algorithms to the hardware and optimized them, brought the unit through lab qualification and supported the flight qualification.

In order to keep all of our team members and our customer developing in parallel, Mercury carefully selected and designed both PC-based and lab-based surrogates for the final hardware – this not only reduced the overall program-level development risk, but also reduced the overall program costs (as opposed to using fully-rugged systems with all development members). The resulting solution met all of the QRC requirements, and was deployed on the battlefield within months of the program start. The eventual success of the program led to increased demand for many more service hours for the platform/sensor suite.

Get Information