- March 26, 2019
March 26, 2019 - Abaco announced the Obox, an R&D evaluation platform designed to reduce the time, expense and risk of developing autonomous platforms for military applications on land, in the air and on the ocean.
A rugged (-40°C to + 71°C) enclosure houses customer-specified combinations of Abaco’s 3U VPX single board computers, FPGA/DSP subsystems and AI/GPGPU-enabled platforms to deliver up to 14 TeraFLOPS of computing power, together with high speed network switches. The Obox has been specifically designed to capture the input from sensors – including Lidar, radar and cameras, integrated over Ethernet – and to drive actuators controlling velocity, trajectory and more. It enables the range of autonomous operations, including detection, recognition, tracking, fusion, perception, planning, decision and effecting.
The Obox leverages COTS (commercial off-the-shelf) technologies, including Xeon processors from Intel; CUDA-enabled GPUs from NVIDIA; and FPGA technology from Xilinx.
Software integration at the ground- and node layers offers the customers pre-consolidation of the image processing and neural network algorithms widely used in machine learning for autonomous vehicle applications.
Two ‘accelerator’ processors support the ability to run neural network-based inference engines to perform tasks such as object detection and image segmentation of multiple video image streams and Lidar point clouds. Developers can utilize GPUs via NVIDIA’s Deep Learning SDK, or FPGAs via Xilinx’s ML suite.
The Obox is fully supported by Abaco’s AXIS software development environment. Of note is AXIS ImageFlex, which provides visualization capabilities to combine video, Lidar, radar and other sensor information into one unified viewer application. It provides high performance algorithms for image transformation to correct for optical distortions in real-time and also to project the image for different viewpoints.
The OBox can optionally be leased, ensuring that it is regularly updated by Abaco with the latest revisions of hardware and software.