By Matthew Elwell, Project Engineer
As the forefront of the development of artificial intelligence and a key application within One Stop Systems’ “AI Transportable” industry, autonomous vehicles present an opportunity for innovation in designing ruggedized servers and high-capacity storage products. While none of the vehicles on the road today are truly self-driving, progress is being steadily made towards systems that can successfully predict and navigate everyday traffic. Autonomous cars and trucks being currently deployed are capable of driving with limited operator input; it is predicted that fully autonomous vehicles will be widely available by the end of the decade. One Stop Systems has had the unique opportunity to support this progression by developing products catered to the extensive storage needs of the autonomous vehicles industry.
With the autonomy of these vehicles continuing to increase, the amount of sensor data required to develop self-driving algorithms will continue to increase as well. Hundreds of sensors are needed throughout the chassis of the vehicle to collect data on its positioning, as well as the positioning of any obstacles around it. The primary types of data recorded are radar, lidar, and video, which are predicted to require bandwidths reaching up to 1024 GB/s by the year 2024. Some autonomous driving companies have circumvented this immense infrastructure requirement by only recording data preceding any “anomalies” experienced while travelling, which usually requires the intervention of a human operator. As for industry standards, the NHTSA only requires the reporting of data leading up to a collision involving an autonomous vehicle. In an industry in which a single failure could lead to severe property damage and loss of life, many companies have opted to store a complete array of sensor data on every trip taken by their vehicles.
As a company focused on deploying servers and storage expansion systems in rugged environments, One Stops Systems has developed a variety of products refined for use in the automotive vehicles industry.
The 3U Short Depth Server (SDS) offers a high degree of flexibility for a variety of AI applications, offering both Intel and AMD based motherboards, alongside 4 dual width PCIe slots for GPU or FPGA accelerated computing.
Designed for use in autonomous trucking, the Centauri storage expansion system is the ideal companion for any server tasked with recording sensor data. Centauri’s compact aluminum chassis holds up to 128TB in 8 2.5” NVMe slots. By connecting to the host server via 3m PCIe uplink cables, storage can be disaggregated from a host server that cannot be accessed by an operator. Both systems feature PCIe Gen 4 architecture, dual redundant CRPS power supplies, fan control and temperature monitoring through IPMI software.
In the autonomous vehicles industry, transferring raw data through conventional means of uploading can take a significant amount of time, bogging down the operation of a vehicle. This issue is aggravated in the training phase of deployment, in which developing algorithms can benefit from processing as much data as possible. The 3U SDS and Centauri systems overcome this storage requirement using a proprietary shared storage canister design. Both products include a bulk removable, hot swappable storage canister that can be seamlessly transferred between either system. Long upload times can be avoided by physically removing and replacing a system’s storage. In an ideal use case, a Centauri unit linked to a server within an autonomous vehicle can house storage canisters that are filled and replaced over the course of a trip. Once the trip is completed, the sensor data can be seamlessly transferred to a 3U SDS located in a datacenter and processed.
Click the buttons below to share this blog post!
Comments will be approved before showing up.