AI/Automation

Are Autonomous Vehicles Going Places Anytime Soon?

(This blog post has been written with inputs from Saraswathi Thippaiah, Principal Consultant - Advanced Engineering & Sreekanta Guptha BP, Principal Consultant - Advanced Engineering, Infosys)

In the foreseeable future, autonomous vehicles (AV) - drones and self-driving cars- will be the major mode of transportation, and facilitate access to a number of allied services. For instance, imagine receiving your products within minutes of placing the order. The AV industry is expected to grow to $126.8 billion by 2027.

Autonomous vehicles drive with the aid of the Internet or a wireless network, using in-vehicle technologies such as sensors, radar, LIDAR (light radar), lasers, cameras, and GPS, to steer, navigate and brake. Data generated or accessed by the AVs is usually stored on a cloud platform. As technology gets smarter, autonomous vehicles will play an increasingly important role in industries such as logistics, where they will simplify fleet management, optimize delivery routes and enable asset tracking. One will also find AVs in industries such as mining, retail, healthcare, and hospitality.

Technology is making autonomous vehicles safer

While we have come close to perfecting the hardware of cars - for efficiency, speed, and safety - the software to power them with no human intervention is still evolving. Self-driving cars are yet to become intelligent and responsive to their environment, and ensure safety of passengers. Working on this challenge is Lvl5, a mapping and localization startup that has developed a way to convert enormous amounts of video footage into high-definition 3D maps of road conditions. These maps will be constantly refreshed to reflect the latest road conditions, providing self-driving cars with the information they need to detect and plan their route safely.

While drones and self-driving cars have been developed to sense the presence of humans and other vehicles, they have hit a bump in the road when it comes to detecting large animals like kangaroos, moose, deer, and others. For instance, in the US, 200 people die each year in animal-related accidents. This led Volvo Cars to study the behavior of different animals and develop an animal detection and collision avoidance system.

Human drivers are responsible for 94% of car accidents worldwide, and in the U.S. 35,200 people lost their lives in 2015. To address this situation, Nauto, a Silicon Valley startup is developing software for self-driving vehicles to collect data on driver behavior and is using computer vision and deep learning techniques to improve driver performance and safety. This data is especially relevant as manned vehicles will share the road with AVs for at least another couple of decades, and both will have to learn each other's behavior to ensure safety.

Will AVs be on the road sooner than later?

Autonomous vehicles recently received a shot in the arm when on July 19, 2017, US lawmakers voted to allow automakers to introduce 100,000 vehicles on the road. However, the assumption is that a licensed driver will be at the wheel and the vehicle will have inbuilt fail-safe features.

While aspiring car owners are excited about the possibilities of AVs (including drones) and self-driving cars, truckers are concerned about their jobs. I think Level 4 and 5 AVs - where in driver attention is either not required to ensure safety, or the driver is redundant - are still very much in the distant future. Some predict as much as a couple of decades away. Much data, technology and testing is needed before AVs are built to be responsive to their constantly changing external environment in real-time. For instance, how to respond to debris on the road, changing traffic rules, read blurred road signs, a sudden flat tire, a brake failure, or bad weather. These are events a human driver would be able to respond to intuitively. Level 1, 2 and 3 AVs, which offer varying degrees of driver assistance, are likely to be on the road soon.

The Infosys engineering services is actively engaged in making AVs a reality in the near future, by developing solutions that can be immediately used by our customers.

We also incubate, develop and mature advanced engineering technologies to renew our existing service lines and create new next-gen ones. Our areas of focus include Robotics and Autonomous Vehicles, Industrial IoT, Machine Learning and Artificial Intelligence, Virtual Reality and Augmented Reality and Additive Manufacturing. All these technologies are evolving continuously and converging towards the Fourth Industrial revolution i.e. Industry 4.0.