Peraton Labs has been awarded a $9.3 million contract by the U.S. Defense Advanced Research Projects Agency (DARPA) to develop artificial intelligence (AI) and machine learning technologies for unmanned ground vehicles. The contract, known as the Learning Introspective Control (LINC) project, aims to enable military systems like drones, robots, and ground vehicles to respond effectively to unforeseen events and conditions.
Traditional control systems often struggle when faced with unexpected circumstances. LINC seeks to address this challenge by creating AI and machine learning-based technologies that allow computers to examine and analyze their own decision-making processes in real-time. This will enable military platforms to adapt and update their control laws on the fly, ensuring stability and control even in the face of unpredicted events.
The LINC-equipped platforms will continuously compare the behavior of the system, as measured by onboard sensors, with a learned model of the system. By doing so, they will be able to identify potential dangers or instability and implement updated control laws as needed. This technology will help operators maintain control of military platforms that have suffered damage or have been modified in the field to meet new requirements.
The program will focus on two key technical areas: learning control through onboard sensors and actuators, and communicating situational awareness and guidance to the operator. Learning control will involve cross-sensor data inference to characterize changes in system operation and rapidly recalibrate control parameters. On the other hand, situational awareness and guidance will provide concise and usable information to operators about changes in system behavior and safety limitations.
The initial phase of the LINC program will involve testing on an iRobot PackBot, combined with a remote 24-core processor. The PackBot is a versatile ground robot weighing 20 pounds, equipped with tracked and untracked flippers, and capable of operating in extreme temperatures. The remote processor, featuring an Nvidia Jetson TX2 general-purpose graphics processing unit (GPGPU) and other cutting-edge components, will support the autonomous decision-making capabilities of the platform.
One of the primary goals of the LINC program is to establish an open-standards-based, plug-and-play architecture that allows for easy integration and interoperability between different software and hardware components. This approach will enable quick additions, removals, substitutions, and modifications of components, ensuring flexibility and adaptability to future technologies.
The LINC project is a four-year, three-phase program that aims to revolutionize the capabilities of unmanned ground vehicles, ships, drone swarms, and robots in responding to unforeseen circumstances. By leveraging artificial intelligence and machine learning, these platforms will be able to adapt and make decisions in real-time, enhancing their effectiveness and resilience in challenging environments.
With this contract, Peraton Labs will play a critical role in advancing the capabilities of military systems, ultimately improving the performance and safety of unmanned ground vehicles and other autonomous platforms. The development of LINC technologies brings us closer to a future where AI and machine learning enable military systems to respond effectively to unpredictable events, contributing to enhanced operational efficiency and mission success.
