Defense Advanced Research Projects AgencyTagged Content List

Automation Technologies

Automatic mechanical or digital operation

Showing 65 results for Automation RSS
Current AI systems excel at tasks defined by rigid rules – such as mastering the board games Go and chess with proficiency surpassing world-class human players. However, AI systems aren’t very good at adapting to constantly changing conditions commonly faced by troops in the real world – from reacting to an adversary’s surprise actions, to fluctuating weather, to operating in unfamiliar terrain. For AI systems to effectively partner with humans across a spectrum of military applications, intelligent machines need to graduate from closed-world problem solving within confined boundaries to open-world challenges characterized by fluid and novel situations.
The current generation of machine learning (ML) systems would not have been possible without significant computing advances made over the past few decades. The development of the graphics-processing unit (GPU) was critical to the advancement of ML as it provided new levels of compute power needed for ML systems to process and train on large data sets. As the field of artificial intelligence looks towards advancing beyond today’s ML capabilities, pushing into the realms of “learning” in real-time, new levels of computing are required.
Military systems are increasingly using software to support functionality, new capabilities, and beyond. Before a new piece of software can be deployed within a system however, its functional safety and compliance with certain standards must be verified and ultimately receive certification. As the rapid rate of software usage continues to grow, it is becoming exceedingly difficult to assure that all software considered for military use is coded correctly and then tested, verified, and documented appropriately.
The Department of Defense (DoD) increasingly relies on software systems to deliver needed functionality, capabilities, and security. However, the rapid pace of software innovation, evolving regulatory requirements, an ever-growing need for stronger system security, and other factors require continual updating and modernization efforts. These produce untenable increases in system complexity and shift the bulk of system costs and developer focus from design and development to maintenance. As this trend continues, the cost and effort required to maintain current systems might constrain DoD’s ability to develop new software-based capabilities.
Department of Defense (DOD) systems and platforms are composed of numerous integrated cyber-physical subsystems, which create an enormous amount of complexity and makes their engineering a daunting task. Today, designing cyber-physical systems (CPS) requires an army of skilled engineers with the right domain expertise, and hundreds of domain-specific tools. The process used to design these systems is largely manual, creating long design cycles that often result in costly redesigns after building and testing the systems. The flaws in the process are numerous – from balancing predictability with cost-efficiency to operating under tight time constraints to integrating disparate pieces from multiple design teams.