Radical Technology in Autonomous Vehicles
May 7, 2020
Imagination Becomes Reality
In today’s world of innovation and technology, what were once ambitious creations of imagination are quickly becoming the new reality, with autonomous vehicles among the latest manifestations. It began with semi-autonomous hybrid and electric cars with power charging and self-parking features and has grown to include “driverless” cars that can navigate on their own. Self-driving technology is becoming increasingly more common and stands to radically transform our transportation system.
Impacts of Autonomous Vehicles
The costs, risks, and benefits of self-driving cars with respect to safety, equity, and the environment continue to be assessed.
- Safety is an overarching concern given tens of thousands of people die in motor vehicles crashes every year in the United States. Theoretically, self-driving vehicles could reduce that number if autonomous artificial intelligence (AI) algorithms can prove to be less error-prone than humans.
- Equity is another major concern. Self-driving technology could help mobilize individuals who are unable to drive themselves, such as the elderly or disabled. On the flip side, widespread adoption of autonomous vehicles could potentially displace millions employed as drivers as the demand for public transportation decreases.
- Environmental impacts are a major uncertainty. Accessible, affordable, and convenient self-driving cars could increase the total number of miles driven each year. If those vehicles are powered by fossil fuels, transportation-related climate emissions could increase. However, if the vehicles are primarily electric based and/or the mass introduction of autonomous vehicles enables a new generation of ‘shared’ vehicles, emissions could potentially be significantly lowered.
How Does a Vehicle Self-Drive?
A complex network of sensors, control systems, computers, and software are used to control, navigate, and drive the vehicle. These self-driving systems require a high-density server or computer located within the vehicle to employ software that processes vast amounts of data from a wide array of sensors like radar, high-powered cameras, and sonar to create and maintain an internal map of the surrounding area moment by moment. The software processes those inputs, plots a path, and sends instructions to the vehicle’s “actuators,” which control acceleration, braking, and steering. Hard-coded rules, obstacle-avoidance algorithms, predictive modeling, and “smart” object discrimination (e.g., knowing the difference between a bicycle and a motorcycle) help the software follow traffic rules and navigate obstacles.
Servers deployed in autonomous vehicle applications are required to have high-performing capabilities in advanced graphics processing, radar, and AI to make crucial split-second decisions on the road. In addition to being powerful, these servers must be small enough to fit within the vehicle’s tight spaces while still being rugged enough to withstand heavy vibration, shock, dust, moisture, and temperature extremes. Reliability and safety are top priorities when traveling from point A to point B.
On the whole, the prospective benefits of autonomous vehicles far outweigh the risks, provided they can be proven safe and reliable in all operating conditions. Of course, that is no easy task given one small error can lead to catastrophic damage. As the technology matures, one of the most essential components of safe autonomous vehicle operation will be to have a rugged, reliable, and high-performing server supporting the demanding and time-sensitive processing requirements.
Mercury Systems’ HDversa server is a comprehensive solution that enables self-driving technology in autonomous vehicles and takes into consideration current and future requirements. Powered by 12 Intel® Xeon® D processors and multiple GPUs, this compact, rugged server easily fits in the trunk of a car. To support scalability and future upgrades, HDversa employs versatile “plug-and-pull” modules that can be quickly modified and upgraded independently or together as-needed. This modular design helps reduce costs by eliminating the need to replace the entire server when implementing technology upgrades or troubleshooting operational challenges. Configuration flexibility simplifies logistics and allows users to develop a solution best suited for the application to achieve optimum performance and precise functionality. The low-power HDversa provides the high-density compute power to execute AI and machine-learning algorithms reliably, helping autonomous vehicles get to their destinations safe and sound.