Nvidia launched the Jetson TX2 series single-card computing module in March 2021, with an ARM processor and GPU. According to the latest report, Aitech, a company that specializes in manufacturing military and space-related components, recently launched a computing module called S-A1760 Venus. The product uses NVIDIA’s single-card Jetson TX2i computing module to help satellites and other spacecraft to achieve high-performance AI computing and video encoding.
The design background of this module is the current growth in demand for advanced imaging and data processing by small satellites. The products can be used in low-Earth orbit and low-Earth orbit satellites and used to access space cameras and other equipment, which greatly reduces the manufacturing cost of satellites.
The NVIDIA Jetson TX2i module is very compact, with Tegra X2 SoC in the center. This processor integrates a dual-core NVIDIA Denver 2 64-bit CPU and a quad-core Arm® Cortex®-A57 MPCore composite processor. It also has a built-in Pascal architecture GPU and has 256 CUDA cores. The module has a built-in 8GB LPDDR4 ECC memory and a 32GB eMMC 5.1 memory chip. The AI performance can reach 1.26 TFLOPS.
Although the Nvidia Tegra X2 SoC chip does not adopt a radiation-resistant design, the Aitech S-A1760 Venus finished module has passed the Series 300 level certification, which can ensure that the product can resist space radiation.
In terms of functions, this computing module of Nvidia supports up to 6 cameras and has 4K 60fps video encoding capabilities. In addition, under H.265 encoding, it supports up to 4 channels of 1080P 60fps video encoding at the same time; under H.264 encoding, it supports 7 channels of 1080P 60fps video encoding. The module power is 10W, 20W optional has multiple PCIe Gen 2 buses, and also has a Gigabit Ethernet connection channel.
