Neousys’ industrial-grade GPU computing platforms are perfect controllers for autonomous robot and vehicle applications
The COVID-19 pandemic has affected our everyday life in many ways. As the pandemic continues to spread around the globe, stay at home and urgent orders to shutdown non-essential businesses have been issued. The remaining are considered to be essential/critical businesses that play a part in our everyday commodities. Along the quiet, lonely streets and under the dimmed street lights, essential workers continue to deliver medical goods and other commodities such as food to people’s doors. They have become a vital link in our everyday lives that have suddenly become, essential.
In addition to current social distancing and minimizing contact restrictions to keep people safe, a more important concern is how to keep essential workers safer. The automation of last-mile delivery is thus considered as an ideal solution, such as automated guided vehicles delivering medication, food to patients in hospitals and delivering packages to people’s front doors. There are also level 2/3 automated minibuses at airports transporting people returning home, coming out of the epicenter, these automated vehicles not only deliver goods but more importantly, they can operate without manpower to eliminate secondary infections to otherwise healthy first-line healthcare providers, bus drivers, carriers, etc.
These automated machines share similar technologies developed for level 4 autonomous driving. They use multiple sensors such as lidar, camera, mmWave radar and ultrasound sensor, to perform environment perception on GPU, and executing route planning accordingly on CPU. With a formula for success, Neousys’ industrial-grade platforms such as 7166GC and POC-500 series are perfect controllers for autonomous robot and autonomous vehicle applications. The Nuvo-7166GC is an NVIDIA Tesla T4 qualified server featuring the latest Intel 9th -Gen Core CPU and Tesla T4 inference accelerator while POC-500 is an ultra-compact fanless controller measuring just 64 x 116 x 176 mm. The systems feature wide range 8-35V DC power input, M.2 2280 NVMe/SATA SSD support, GbE (Nuvo-7166GC)/PoE (POC-500), USB3.1 and COM ports for connecting sensors and communicating with other devices. In addition, the systems feature mini-PCIe and M.2 slots for wireless module expansion so the automated machines can always stay with cloud servers. With the amazing inference capability brought by NVIDIA Tesla T4, Nuvo-7166GC is the perfect brain for machines to see and perceive the environment, while the POC-515 aside, possessing AMD Ryzen™ 4C/8T CPU, can be the second brain that plans the route and communicates with others.
With deep learning and AI inference technology, Neousys GPU platforms have been success stories worldwide, such as the central processing unit for automated vehicles in the open-source Baidu Apollo. The systems offer other advantages such as IP rating protection and 3Grms vibration resistance. The unique thermal solution design is one of the first in the industry to allow for the integration of GPU processing power. Neousys systems simplify the heat conduction path and dissipate heat by placing extremely efficient thermal interface materials on the CPU and electronic components, allowing them to directly make contact with the external heatsink and dramatically boost its capabilities beyond standard IPCs.