Cloud-Based Efficient Data Analysis and Reduction (CEDAR)

Architecture Technology Corporation (ATCorp) proposes the Cloud-based Efficient Data Analysis and Reduction (CEDAR) system, a novel hardware and software solution to efficiently increase and manage shared compute resources located within an autonomous vehicle to run mission-specific software that analyzes, combines, and filters sensor data in real-time or near real-time according to changing needs over the course of a mission, autonomously or under the control of mission commanders. CEDAR allows a mix of networked low power and medium power compute platforms with varying capabilities including but not limited to GPU support, parallel compute, and FPGA programmable capabilities to be dynamically provisioned among multiple software programs in isolated security enclaves to address competing sensor data analysis needs. It does this while providing fast, direct communication of sensor data, usually without the need for cross-domain guards, and without compromising security enclaves. The compute mix can be upgraded or modified as analysis algorithm needs changes or as available compute platforms advance.