According to industry analyst forecasts, $1.5 billion worth of processors specifically designed for autonomous driving will be needed by 2023 when 7.3 million vehicles (or 7% of the market) would have this capability. When more than 50% of all sold vehicles are categorized as SAE Level 3 or higher, as specified by the National Highway Traffic Safety Administration, this is anticipated to increase to $14 billion in 2030. (NHTSA).
In spite of the fact that photonic chips are quicker and more energy-efficient, they will need fewer chips to achieve SAE Level 3; yet, we may anticipate that this improved compute performance will hasten the creation and accessibility of completely SAE Level 5 autonomous cars. If so, the projected $14 billion market for autonomous driving photonic processors by 2030 is likely to be much exceeded.
When you take into account all of the diverse potential uses of autonomous electric vehicles (AEVs), such as taxis and service vehicles in major cities or the clean transportation of goods on our highways, you can see how this technology can quickly start to significantly impact our environment: by bringing clean air to some of the most populated and polluted cities.
AEVs must use a bewildering variety of sensors, including cameras, lidar, radar, and ultrasonic sensors, to function effectively and securely. Together, they gather information to detect, respond, and forecast in real-time, effectively acting as the vehicle’s “eyes.”
The ability to offer autonomy while still retaining battery range is a challenge for autonomous electric vehicles (EVs). While a short- or even medium-term solution to the AEV issue using quantum computers is doubtful, instead of using electrical impulses to compute and convey data, photonic computers employ laser light. As a result, the power consumption is drastically reduced, and crucial processor performance-related factors like clock speed and latency are improved.
Additionally, photonic computers allow inputs from several sensors to execute inference tasks simultaneously on a single processor core (each input encoded in a different color), whereas a conventional processor can only handle one task at a time.
Hybrid photonic semiconductors provide an edge over traditional systems due to the unique characteristics of light. The same neural network model is used for all data inputs, each of which is represented in a distinct wavelength, or color. As a result, photonic processors not only have higher throughput than their electronic equivalents but also use a great deal less energy.
Applications like cloud computing and, perhaps, autonomous driving, which need the real-time processing of enormous volumes of data, benefit greatly from photonic computers’ tremendous throughput, low latency, and very low power requirements.
As photonic computing technology nears commercial availability, it has the potential to accelerate the current autonomous driving roadmap while simultaneously lowering its carbon impact.
Source:- Techcrunch