Alexandr Khomich, Founder of Andersen.
According to the fresh data obtained by Boston Consulting Group during their global-scale research, an impressive 29% of the value from all artificial intelligence (AI) initiatives is driven in the automotive industry. At the same time, the research states that a mere 26% of businesses have implemented the strategies and infrastructure needed for their AI projects to generate value.
In this article, I would like to describe some practical applications of AI in automotive that carmakers, OEMs and technology companies can leverage today and that mark the future of the industry.
Today, car manufacturers can use AI in two main areas. Firstly, they can rely on industrial robots in factories capable of making real-time decisions that will enhance productivity. Among Andersen’s offers that support the manufacturing process are quality control mechanisms driven by computer vision, predictive maintenance tools, software for making projections of supply and demand and digital twins.
Secondly, AI has been widely implemented in vehicles themselves for purposes including navigation control. Below, I’ll explore the areas where this technology already yields substantial benefits.
AI And Machine Learning Controlling The ADAS
Today, vehicles rely on LiDAR to control advanced driver-assistance systems, also known as ADAS. However, LiDAR has limited capabilities. If a driver uses adaptive cruise control, LiDAR sensors can regulate the distance between their vehicle and the vehicle in front of them, but not much more than that.
There are car models already where AI fully controls the ADAS, such as MONA M03 by XPENG, a prominent Chinese OEM. MONA stands for “made of new AI.” How does this car operate without LiDAR?
In China, 3-D prints of cities are drawn based on data from surveillance cameras, mapping apps and sensors integrated into city infrastructures. These prints include the grids of sidewalks, traffic lights and other objects that AI systems continuously learn from. Smart systems embedded in cars also learn from each other and from their previous rides around the city, which leads to a safer environment on roads. Thus, AI can instantaneously make better decisions than the driver.
In our recent interview with Maxence Van Goethem, fleet sales manager for the Benelux region at XPENG, published on the Andersen RnD YouTube channelhe shared his opinion on this matter: “In five to 10 years, LiDAR will be replaced by AI and machine learning.” In other words, the generation after the vehicles that have already been designed and are currently being tested will rather use AI than re-implement LiDAR systems.
AI Enhancing The Work Of V2X, V2C And V2P Systems
Vehicle-to-car (V2C) and vehicle-to-pedestrian (V2P) concepts are, in essence, subsets of the vehicle-to-everything (V2X) framework, which paves the way to autonomous driving systems that boast increased safety. According to the National Highway Traffic Safety Administration, V2X can reduce crashes by 615,000 accidents per year.
How can AI improve the work of the technologies related to the V2X, V2C and V2P concepts? In the case of the V2C, the cars embedded with intelligent systems will communicate with each other. For example, they will send alerts about a dangerous situation on the road so that other cars would be aware of it several kilometers in advance, before they’re in the vicinity of that place of accident, road repair and so on. The AI-fueled software in cars will then regulate the work of ADAS so that vehicles can handle these hazardous situations.
V2P, in turn, is a crucial part of an ADAS, but at the same time, it’s challenging to implement. This is because it’s easier for a camera or a sensor to spot a metallic sheet or a wall than a moving object like a person or an animal. However, ensuring the safety of pedestrians is critical, and this is where AI can help enhance the capabilities of V2P systems.
In China, a unique strategy for autonomous vehicle deployment has been in place over the last few years. It implies vehicle-road-cloud integration (C-V2X), meaning that vehicles can exchange data with roadside infrastructures through cloud computing.
An eloquent example of this strategy in action is the establishment of the Yizhuang Pilot Zone in the Chinese capital. More than 500 self-driven vehicles and 1,000 microcars operate within a grid of smart intersections equipped with AI-enabled cameras and edge computing. The initiative is backed by local and central governments, and the major technology developers and providers are giants like Baidu, WeRide and others.
The innovation is expected to improve safety on the road by empowering vehicles to detect unseen traffic around intersections, notice pedestrians ahead of time and trigger emergency braking when needed.
How Can We Accelerate The Adoption Of Smart Technology In Vehicles?
Today, it’s still up to the human behind the wheel to make decisions. However, the future is nearing when the driving will be entrusted to machines that are constantly learning and communicating with each other. In a few years from now, technology and AI will be so much integrated into our driving routines that it will be difficult to think of a time when it wasn’t so.
What do European carmakers need to keep pace with Chinese innovation? They will have to switch from their legacy ways of working that they’ve been implementing over the last hundred years or so and follow the example of Chinese OEMs who continuously look for advancements in technology to gain the upper hand. As we see the advancements coming in from China, we’ll also see the major European manufacturers like BMW, the Mercedes-Benz Group, Volkswagen Group and more adopt AI and retrain their workers.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
