By Myles Hosford,
Everyone is in the race to perfect the autonomous car, developing a self-driving vehicle that is safer, smarter and more efficient. To develop and deploy autonomous cars at scale, the right technology infrastructure is critical as the production of these vehicles requires high-performance computing capacity, along with seamless cybersecurity, and the ability to manage vast data sets. The cloud offers organizations the support they need to realise their dream of building their self-driving vehicle.
Training and testing
Autonomous cars strive to be safer than vehicles operated by humans. Delivering on this ambition requires extensive modelling and testing. The ability to collect, store, and manage data is critical, as are advanced machine learning techniques.
Toyota Research Institute (TRI) believes that accurately training autonomous cars requires trillions of miles of testing. To deliver on this, it has a fleet of test cars equipped with Light Direction and Ranging (Lidar) Sensors to record data, collecting terabytes of data every day, needing quick retrieval and analyses. TRI uses AWS to manage this data and access the processing power required to train machine learning models quickly. TRI now retrains vehicle models, increases accuracy, and introduces new features faster thanks to the cloud’s infrastructure. By following similar models, more automotive businesses will accelerate the development of safer cars.
Enabling autonomous cars to make rapid, data-driven decisions will make our roads safer. These machines need backing by reliable infrastructure with low latency and high availability. They also need to analyse information in real-time, including data on road conditions, weather, and the behaviour of other vehicles. Applying AI allows the car to react swiftly and safely to road conditions.
When a second of lag can make the difference between a safe or dangerous response, autonomous vehicles do not have the luxury of waiting for data processing in the cloud. Edge computing allows the analysis of critical data in the car and in real-time which reduces the cost of transmitting additional data to the cloud.
Security in the cloud
Cybersecurity is important to protect autonomous cars from hackers and malware to ensure they cannot gain access to driving controls or the data that runs through each vehicle.
It’s the job of the cloud provider to ensure cybersecurity updates and upgrades regularly happen. Additionally, the best cloud providers deliver automated security services which apply machine learning to proactively manage tasks including security assessments, threat detection, and policy management. By having security baked in, manufacturers can be confident that they have the solutions in place to detect new and emerging vulnerabilities and threats which will reduce harm to drivers and lower the risk of a breach.
Connected cars, cloud and security
Autonomous cars are our future; however, to drive adoption, manufacturers must ensure they are secure and supported by robust infrastructure. Using the right cloud vendors is critical, enabling organisations to focus their resources on building differentiated automotive experiences, rather than managing IT infrastructure.
With this in mind, AWS provides a full suite of services including Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicle development and deployment. AWS’ nearly unlimited storage and compute capacity and support for deep learning frameworks such as Apache MXNet, TensorFlow and PyTorch accelerate algorithm training and testing. AWS Greengrass provides edge computing with machine learning inference capabilities for real-time processing of local rules and events in the vehicle while minimizing the cost of transmitting data to the cloud.
A combination of scalable storage and compute capacity and support for deep learning frameworks help accelerate testing and service development. At the same time, our agile platform helps businesses enables greater innovation, improve security posture and lower IT cost structure bringing us all one step closer to a reality where autonomous vehicles are the norm.