The 5 Most Amazing AI Advances in Autonomous Driving

AutoFacets

, Machine Learning & AI

Takeaway: Artificial intelligence is an integral component in autonomous vehicles, and the reason behind recent technological advancements.

 

The very idea of a driverless vehicle rolling around on the streets seems incredible. And yet, we may be close to seeing such vehicles on the road around the world, thanks to artificial intelligence (AI), among other driving forces. In the recent past, there have been some amazing advances in autonomous vehicle technology which indicate the dream is inching toward fruition. It seems that the framework of autonomous vehicles has been almost finalized. Subject to legal and administrative approvals, driverless vehicles will be a common sight on the roads soon. (To learn about other automotive advancements, check out 5 Ways Our Cars Have Become Computers.)

Delivery Vehicles

You have seen delivery vehicles driven by humans delivering packages. Now, we could see the same task done by driverless vehicles – and with higher efficiency and swiftness. Nvidia, the leading computer graphics provider, Deutsche Post DHL Group (DPDHL), the world’s biggest mail and logistics company, and ZF, an automotive provider have teamed up to deploy driverless electric light trucks which will transport and deliver packages. The driverless trucks will deliver packages from a central point to the destination. In the interim, it is trained to accurately assess its environment for variables such as traffic conditions, parking spot identification and parking, and pedestrian behavior. The truck is powered by the ZF ProAI self-driving system, which is powered by the Nvidia DRIVE PX palm-size supercomputer, but it also includes sensors, cameras, LIDAR and radar that feed the data into the system. Note that apart from the obvious benefit of relentless accuracy and no driver fatigue that the technology promises, there is also the potential of huge cost savings because the process of delivering packages from the central point to the destination is the most expensive for logistics companies.

Full Autonomy

Imagine luxurious driverless taxis that help passengers move between points. You can just do your thing – watch a movie, work on your laptop or listen to music – and not have to worry about the taxi safely taking you to your destination. Such taxis could soon become a reality. Nvidia’s DRIVE PX AI platform is going to usher in fully autonomous vehicles. The DRIVE PX AI platform is 10 times superior to its predecessor DRIVE PX 2 and can handle over 320 trillion operations per second. This means the car will be learning and making accurate decisions about its environment on the road much faster than its predecessors.

Currently, Tesla cars are equipped with the necessary hardware for autonomous driving, but software updates are required to fully enable the feature. While it will allow fully autonomous driving, it will also still allow the human driver to take control, when necessary. The next generation of autonomous vehicles would not need steering wheels, pedals or transmissions. Such cars will potentially reduce accidents, will be viable transportation options for the elderly or those with vision or physical disabilities, and could increase productivity.

Parking

Car parking is not really a novel development. The advent of automated parallel parking is probably among the earliest of AI exploits in autonomous driving technology. However, the concept has greatly evolved in recent years. Parking, especially in big cities, is a major problem because it increases emissions, wastes time and productivity, and increases stress. Bosch has developed a smart AI-based system that provides data on available parking spots, locations and times to park. The car even does the parking itself without any accidents. As the car is on the move, it receives information about parking availability at places closest to its GPS location. The parking space data is sent to many cloud servers from cars, which is then sent back to cars so that drivers can learn about parking space availability.

Cars with Common Sense

While work on the autonomous driving domain has seen amazing advancements already, common sense like that of human drivers has been the missing piece in the developments. In difficult traffic conditions, especially in big and chaotic cities, the human mind is highly sensitive to constantly changing variables such as fellow drivers’ attitudes, pedestrian behavior and erratic weather. It is critical for the driverless car to develop a human-like common sense on the streets. An MIT spinoff, known as iSee, has been working on AI and deep learning to impart common sense into driverless cars. This is going to be the most significant component of the autonomous vehicle initiative. The iSee team has been working hard on data and neural networks so that cars can learn from data and negotiate any and all types of traffic conditions. According to Yibiao Zhao, co-founder of iSee, “The human mind is super-sensitive to physics and social cues. Current AI is relatively limited in those domains, and we think that is actually the missing piece in driving.” (For more on deep learning, see A Tour of Deep Learning Models.)

Cars with Peripheral Vision

Knowledge of pedestrians, objects or vehicles around a blind corner is a critical factor in safe driving. Blind spots have been responsible for many accidents. A new AI technology enables cars to view and assess the distance and speed of pedestrians, objects or vehicles around a blind corner. CornerCameras, an AI initiative by MIT researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) enables driverless cars to identify people or objects situated in blind corners of the roads. The technology uses light reflections and does not actually see objects or people. From the data received, it can direct the self-driving car for a better driving experience. According to Katherine Bouman, the lead author of the paper detailing the system, “Even though those objects aren’t actually visible to the camera, we can look at how their movements affect the penumbra to determine where they are and where they’re going.”

Conclusion

These developments are exciting news and are expediting the arrival of the fully autonomous car. However, before we see autonomous cars on the road across the globe and it is treated as a normal phenomenon, two things will be key: one, the imparting of common sense in driverless cars, and two, clearance of the various legal and insurance hurdles on the way.


Courtesy of Techopedia 

 

Leave a Reply