White Paper

Enabling edge processing for military intelligent sensors with stacked, high-speed DDR4 and DDR5 memory

Issue link: https://read.uberflip.com/i/1173434

Contents of this Issue

Navigation

Page 0 of 3

w w w. m r c y. c o m WHITE PAPER Enabling edge processing for military intelligent sensors with stacked, high-speed DDR4 and DDR5 memory An unmanned aerial vehicle (UAV) on a pre-defined border patrol route over an unstable territory collects and analyzes intelligence, surveillance and reconnaissance (ISR) data for movements of an adversarial camp. Programmed to transmit alerts upon certain movements on the camp and surrounding areas to the tactical command center and soldiers in formation, the onboard artificial intelligence (AI) system enables the ground unit to focus on their mission plans. The processing system developed for high-speed performance using a foreign manufactured 4GB double-data rate fourth generation (DDR4) stacked memory device is not running at the specified speed; signal interference within the die stack has compromised data integrity. The error correction code (ECC) within the algorithms filters out the corrupt data resulting in no data transmission. The system provides no alert to the ground forces on enemy movement towards the soldiers positioned closest to the camp. The soldiers have no idea the enemy is approaching. In military environments, seconds can be the difference between mission success or failure and life or death. System malfunction or failure due to poor signal integrity within integrated circuits and embedded board designs leads to catastrophic events. A soldier in hostile territory relies on accurate, real-time data from a variety of sensor systems, including those in unmanned vehicles, to accurately analyze threats and take swift action. These systems are small enough to be deployable in the battlefield, yet do the work of large data centers from brick and mortar command centers. AI and autonomous systems allow machines to perform operations previously only accomplished by a war fighter. ISR missions with manned aircraft are now performed with UAVs and unmanned ground vehicles (UGV) with integrated AI. Independent sensor systems, now replacing soldiers on foot patrol, are installed at a base camp to collect data, analyze it and warn of incoming threats. In all cases, these systems increase the efficiency and security of our warfighters while improving situational awareness. These systems acquire, digitize, process, store and disseminate data into action based on sophisticated algorithms processing sensor data, comparing patterns, performing predictive analyses and reacting before a warfighter sees a thing. This real-time processing of data at the point of generation and consumption, decentralized from a data center or the cloud, is edge processing. These edge processing architectures require real-time and high performance processing systems with low latency. JENNIFER KEENAN, SR. PRODUCT MARKETING MANAGER, MICROELECTRONICS SECURE SOLUTIONS GROUP SEANN AYERS, SR. PRINCIPAL SIGNAL INTEGRITY ENGINEER, MICROELECTRONICS SECURE SOLUTIONS GROUP APRIL 2019

Articles in this issue

view archives of White Paper - Enabling edge processing for military intelligent sensors with stacked, high-speed DDR4 and DDR5 memory