Whilst AI in the automotive industry is currently in its fledgling stages, it has the potential to completely revolutionise the sector. AI is being implemented, not just in headline-grabbing self-driving vehicles but across the entire value chain from manufacturing and design through to aftermarket services such as predictive maintenance, personalisation and insurance.
How artificial intelligence is used in the automotive industry
The race is on to get fully-automated vehicles (AVs) on the world’s road. Established car manufacturers and tech start-ups are all working on the technology that, if successful, will have far-reaching impacts across the automotive industry and beyond. Manufacturers are not just seeking to replace privately-owned vehicles but to completely change the way the world moves from last mile delivery to ride share.
A study by theSouthwest Research Institute claims that AVs can lead to as much as 20% improvement in fuel consumption. A separatestudy by ICCT (The International Council on Clean Transportation) concluded that using electricity as the engine power source for the vehicles could reduce greenhouse gases emissions by as much as 72%.
As well as a reduction in emissions, fully autonomous vehicles could also encourage a reduction in road traffic. If ride-hailing apps were to adopt AVs for instance, this would allow for a car-sharing at all times of the day and, without the need of a driver, vehicles could be designed to accommodate more passengers. Some of the most prevalent IoT technologies and applications that modern vehicles are fortified with include:
- Sensors that gather valuable data about driver behavior and vehicle condition
- Complex machine learning (ML) algorithms that convert the data collected to insightful reports
- Usage of the data collected to segment customers as well as provide individualized offers to customers
Applications of ML and AI: how artificial intelligence is used in cars
Whilst AVs edge ever closer, AI is already playing a vital road in driver assistance technology. Advanced driver assistance systems (ADAS) capabilities such as lane departure warnings, blind spot detection and emergency braking have already achieved widespread market penetration, significantly increasing driver safety and improving the driving experience.
ADAS is enabled by technologies that allow on-board computers to perceive the external environment of the vehicle using an array of onboard sensors and data processing. Cameras, radar, ultrasonic sensors and, in some instances, LiDAR collect vast amounts of data from the vehicle’s surroundings, providing 360° awareness. The aim is to enhance the capabilities of the driver and where necessary assume control of the vehicle to ensure the safety of all road users.
ADAS capabilities range from level 0 which includes informational messages to the driver such as parking sensors and lane departure warnings, through to level 5 which is fully automated with no requirement for a driver.
Current driver assistance tools on the market are designed to increase safety for all passengers in a vehicle. However, even in the most advance ADAS Systems on the market, such as Tesla’s Autopilot, it is the driver that is ultimately responsible for the operation of the vehicle. In a truly autonomous vehicle, a driver is no longer required and this creates an opportunity to step away from the age-old vehicle design. Vehicles can be redesigned to allow for more passengers or to create space to work. In July 2021, Porsche released a concept for the interior of a self-drive car. The vehicle, designed to seat up to six people, features front and rear sliding doors and a ‘driver’s’ seat that rotates 180 degrees, so they can turn to face fellow passengers when the vehicle is in autonomous mode.
Modern vehicles continue to grow ever-more complex and there is an increasing demand for cost-efficient technical solutions to ensure the vehicles’ functional safety and reliability over its lifetime. AI, and particularly machine learning, plays an essential role in predictive maintenance.
New cars are fitted with a range of sensors and other instruments that generate data and provide insights into the operation of the vehicle. According to the European Data Protection Supervisor, a new car typically has over 100 sensors, and generates as much as 25 gigabytes a day. Machine Learning models fetch sensor data from the vehicle to track the critical components like brakes, battery, and engine. Proactive measures can then be taken to avoid breakdowns, saving the customer time and money.
Consumer data analysis
As well as providing information on the workings of the vehicle, connected cars also generate vast sums of data on the driver. The value of data generated by vehicles has been recognised for some time, not least by Big Tech companies which are increasingly integrating themselves into automotive supply chains. Apple for example allows iPhone users to connect their phone to their car’s “infotainment” systems so that it covers more of the vehicle’s control functions and BMW driver can use their phone in place of the car’s electronic key.
Advertisers can make use of this data to deliver timely, targeted messages towards the customer. Sensor data from systems like OnStar or vehicle plugins like Snapshot from Progressive provide data that can improve customer experience by offering monetary incentives to customers who drive safely. Data captured in these ways can also be used to monitor traffic trends and help local governments improve traffic management and infrastructure, such as filling in potholes around the city.
Predictions: the future of AI in automotive
AI is already being applied in many ways, from the fantastic to the necessary. Ask anyone about the future of AI in the automotive world, and they will say self-driving cars, and to an extent, they’d be right. But the world of self-driving cars is a tough one, with many hurdles to overcome first. Ignoring the ethical constraints, there is also the issue of the fact that no two cities are alike. Sure, introducing autonomous vehicles to grid-system cities is fine, but Europe doesn’t have that luxury, with ancient cobbled streets and poorly designed traffic systems in some cases.
The fact is, autonomous automotive is a nice long-term dream, but in the short-term, AI is more likely to be used in safety and service applications. Introducing new ways to keep driver, passengers and others safe through the use of AI in ways to stop or recognise potential incidents. Ways to react to crashes whereby the AI could potentially resolve a situation faster than the driver, through driver monitoring and assistance.
Serviceability, longevity and preventative maintenance, is going to be paramount for customers as the rules around fossil fuel car sales come into play. This means the last generation of fossil fuel powered cars will have to last for a longer time than previously. AI will be used in two ways. One, for manufacturers to understand and predict failure points, and create enough inventory to cover a longer service life of the vehicle. Two, to help the customer identify an issue and offer guidance of how it can be resolved.
The future of AI in automotive has to be around sustainability first and foremost. The landscape of the automobile market is changing rapidly, and AI needs to be there to keep up with the pace.
How artificial intelligence is changing the auto industry
Artificial intelligence has been involved in the auto industry for some time, with the automation of production processes via robotics. However, today, it is ever more present in other areas of the industry.
Faster supply chain requirements, smaller inventory stocks, mean that supply chains are becoming more reliant on AI. If anything can be learned from the chip crisis, is that AI is detrimental to working to avoid these supply chain issues in the future.
Car design today relies heavily on AI, to understand wind-resistance, improve performance in hypercars, or even how to reduce fuel and consumption. This then leads into manufacturing, where robotics, materials movement, and supply chain management, is all determined and controlled via artificial intelligence.
But look towards China and you will see another focus. Chinese car manufacturers have had to play catch-up to the rest of the world, and one of those areas is quality. The post-production quality control has become a slightly different scenario from what it was many years ago, when a car was essentially a piece of machinery. Today, it is more a glorified computer. AI is being used to automate and manage the quality control process, to ensure that tests are completed correctly, and that all components are within parameters.
AI is also being used to assist in manual processes. In 2018, Hyundai introduced a wearable robotic exoskeleton which assists workers in their tasks, to help with more complex work, but also reduce fatigue and injuries by up to 80%.
A2i’s PriceCast Fuel has been a prime component of the automotive industry in the form of customer centric fuel pricing. The AI is a self-learning neural-net technology, which automates the understanding of customer’s perception of fuel pricing, and the market conditions surrounding that perception, and sets a price accordingly. This means that opportunities to gain margin and volume share increase, thanks to recognition incremental changes in the market.