Demystifying the essence of 4TB in driverless cars
Author of this article
Kathy Winter
Vice President of Intel Corporation and General Manager of the Driverless Division. She joined Intel from Delphi in 2016. At Delphi, she successfully achieved the first multinational journey of a fully driverless car.
If 3,000 people speak at the same time, can you still hear what everyone is talking about?As an engineer, I am passionate about solving problems and using "mathematical language" - or numbers - to understand our world. Numbers, not just numerical values, add color to the story in ways that text can't. The big numbers are very interesting, because their meaning is usually more complicated than the value itself, especially for 4TB, and I am very excited to think about the significance of this number for the driverless industry.
First of all, why is this number? A driverless car is expected to generate 4 terabytes of data in an hour and a half, and each person spends an average of one and a half hours in the car every day. By 2020, 3,000 Internet users will generate so much data every day. Now, you may not be able to intuitively experience such a large amount of data. Then let's change the angle, how many people have 3,000 friends on Facebook? Imagine that you are paying attention to and receiving new updates from these friends every day.
The amount of data generated per second by the sensors of the driverless car
If the highlight of a driverless car's data is only 4TB, it may not be exciting. What makes data a driverless “new oil†and the real challenge is that we need to use that data to turn them into actionable insights that allow cars to think, learn and drive without human intervention. In the future, data-driven driving is expected to reduce accidents caused by 90% of human error.
Intel is a data company. We understand how to create, move, store, process, analyze, and manage large-scale data. Moreover, we are applying this expertise to the driverless industry. Based on our experience, we know that the fastest way to solve the data challenges faced by driverless is industry collaboration. There is still a lot of work to be done if you want to fully realize driverlessness in 2021, but I believe we can work with industry and partners to achieve this goal.
Driverless data includes three basic types: technical data, crowdsourced data, and personal data. The technical data is the most obvious. The data comes from a set of sensors that help the car "see" the surrounding environment. These data help the car identify people or fire hydrants, "pay attention" to road squats, or calculate the speed of the next car. In addition, technical data helps capture new driving scenarios and pass it to the cloud for learning and improving software that controls driving behavior. When technical data is transmitted to the cloud, it can benefit all vehicles connected to the cloud.
Crowdsourcing data is data that local cars receive from the perimeter, such as traffic conditions and road conditions. You can imagine that many applications can use this type of data, for example, looking for nearby parking lots or avoiding traffic congestion. Finally, personal data, including the radio stations you want to listen to, your favorite coffee shop, your preferred route, and more. This type of data helps create a more engaging and personalized experience in your driverless car.
As the industry moves toward driverless cars, data presents many challenges for the global industry. The first challenge is the 4TB mentioned above. The amount of data is exponentially growing, which requires huge computing power to organize, process, analyze, understand, share, and store. Imagine the computing power of a data center server, not a PC.
Training a driverless car as soon as possible is another challenge. Machine learning, simulation, and algorithm improvements must occur immediately when a new driving response or situation is identified—not weeks or months later, and the updated driving mode must be pushed to the car as soon as it is available. When, where, and how these factors are important are important to the present, especially in the future when unmanned driving becomes the norm.
In addition, data protection and consumer trust in driverless driving are also worthy of attention. I am often asked: How to store and share data securely? We are extremely serious about this, what data is stored? What data is discarded? What data to share? How do we protect this data? These issues require industry collaboration and require our top experts to resolve them.
Finally, the challenge of data is not constant. As the number of driverless cars grows from small to hundreds of millions, the data challenge will increase. This challenge can only be met by being able to handle the sheer volume of data. As supercomputers and the cloud behind them become more sophisticated, true system scalability is critical for the interior of the car—back to 4TB of data—and for large external data centers.
No company can solve these data challenges independently. Intel believes that the best way to address the data challenges of driverless driving is industry collaboration to jointly develop a secure top-level platform and share security-related information. Our common goal is to have no accidents and to facilitate travel for all, and industry cooperation will accelerate the achievement of this goal. I am very excited to be able to work with my partners on the 4TB challenge in the Intel team, because I know that solving this problem will make the road safer and make the trip better.
2.5 Small SATA SSD for PC Laptop
Compatible Devices: Compatible with desktops and laptops that accept 2.5" 7 millimeter SATA drives.
Trustworthy and Reliable:Supports TRIM command, Garbage Collection technology, RAID, and ECC (Error Checking & Correction) to provide the optimized performance and enhanced reliability.
Portable Usb C External Hard Disk,Laptop Hard Drive Enclosure,Hard Disk For Desktop Laptop,120Gb Hard Disk Memory
MICROBITS TECHNOLOGY LIMITED , https://www.hkmicrobits.com