Craig Milo Rogers

FFTW and Complex Ambiguity Function performance on the Maestro processor

Maestro is a 49-core general-purpose, radiation-hardened processor for space. Fourier Transform is an important computation in many space applications. FFTW is a high performance package that computes one and multidimensional Fourier Transforms. We ported FFTW library to the Maestro processor and used it extensively in implementing the Complex Ambiguity Function (CAF), which is a reference space application. We were able to obtain 145 MFLOPS (0.36 FLOPs/Cycle) on a single Maestro core with double-precision computations and 202 MFLOPS (0.50 FLOPs/Cycle) on a single Maestro core with single-precision computations running at 260 MHz.12

Interfaces for understanding multi-agent behavior

>Synchronized punch-card displaysare an interface technique to visualize tens of thousands of variables by encoding their values as color chips in a rectangular array. Our technique ties multiple such displays to a timeline of events enabling the punch-card displays to show animations of the behavior of complex systems. Punch-card displays not only make it easy to understand the high-level behavior of systems, but also enable users to quickly focus on individual variables and on fine-grained time intervals. This paper describes synchronized punch-card displays and shows how this technique is extremely powerful for understanding the behavior of complex multi-agent systems.

Distributed coordination in uncertain multiagent systems

We consider real-time multi-agent coordination in a dynamic and uncertain domain addressing both distributed state information and partial knowledge of the common reward function. The challenge is to find functional strategies when bounded rationality hinders the ability to encompass the values of possible sample paths of the system. This paper discusses a new approach based on assigning agents to monitor portions of the reward structure for which they aggregate and propagate appropriate profiles which compactly represent relevant information used for policy modification. This approach shows promise as an alternate and potentially superior technique with respect to current decision-theoretic and scheduling approaches.

CSC: Criticality-Sensitive Coordination

Our Criticality-Sensitive Coordination (CSC) agents are designed to enhance the performance of a human-team working together in uncertain and dynamic settings by monitoring and adapting their plans as dictated by the evolution of the environment. Such situations model military scenarios such as a coordinated joint operations or enterprise settings such as multiple-project management. Among the many challenges in these situations are the large space of possible states due to uncertainty, the distributed / partial knowledge of current state and plan among the agents and the need to react in a timely manner to events that may not be in the original model. In fact, reaction alone is often insufficient as in environments where success depends on completing sequences of coupled actions, one needs to anticipate future difficulties and enable contingencies to alleviate potential hazards.

The Maestro Flight Experiment: A 49-core radiation hardened processor in space

The Maestro Flight Experiment (MFE) demonstrates the operation of a modern, many-core radiation hardened processor (the Maestro ITC) in space. The MFE executes diagnostics, performs image processing activities, addresses strategies for dealing with thermal and power constraints in orbit, and assesses the sensitivity of the processor to radiation exposure in low earth orbit. The MFE demonstrates a substantial advance in radiation hardened on-orbit computational performance, which is needed to handle the ever-increasing data load of modern imaging sensors and other orbital sensor systems.