The Way AI Leads Innovation In Automotive Sector

By Sherry C Kunjachan:

In a recent research report, IHS, a leading source of critical information, has predicted that Artificial Intelligence (AI) systems for Advanced Driver Assisted Systems (ADAS) and infotainment will increase from 7 million in 2015 to 122 million by 2025. Hence, the implementation of AI in automobiles is expected to rise by 109 percent till 2025 from just eight percent in 2015. With techniques such as sensor fusion and deep learning, researchers have developed a technology that would help build a three-dimensional map of any activity around the car. Hence, in recent developments, automotive companies such as Ford, General Motors and Toyota have invested in collaborating with AI companies. Ford and Toyota have invested about a billion dollars each with Argo AI and Toyota Research Institute (TRI) respectively. General Motors has partnered with IBM to supercharge its OnStar platform with IBM Watson’s supercomputing power.

AI Learning

AI can be considered as a machine’s ability to understand what’s happening around it and act to complete an objective successfully. AI systems could be broadly classified into two categories: weak and strong. The weak AI systems are those designed and trained for a specific task. But the strong AI systems are more similar to human intelligence. In a given and unfamiliar situation, it can find a solution. Machine learning belongs to a type of AI that allows software to become more accurate in its predictions, without being explicitly programmed and can be either supervised or unsupervised. In supervised situations, humans provide input and the desired output, and the systems will apply the ‘learnings’ when new data is presented. In the unsupervised scenario, the system uses an iterative approach to review the data and arrive at conclusions which are also known as ‘Deep Learning.’

AI In Automotive Sector

Considering the latest technology updates, the five primary applications of AI in the automotive sector are:
i. Assisted/Autonomous Driving: The most visible applications of AI is seen on board the car in the form of cruise control systems. As we move from these features to applications that require real-time cogitative capabilities like pedestrian detection, object identification, traffic sign recognition – more robust AI techniques are necessary to achieve the required performance and accuracy objectives. Moreover, as we move towards truly autonomous driving scenarios, the system should adapt to and work well in any scenario that it faces. To achieve the fully autonomous driving capability, the AI systems need to learn from a large amount of real data before it can be put in production.
ii. Automotive manufacturing: With the onset of Industry 4.0, carmakers are also leveraging AI in the factory for reducing chances of equipment failure with manufacturing automation, increasing human-robot collaboration, super-charging the supply chain for hassle-free procurement and finally smarter project management with improved business support functions.
iii. Cloud-hosted Intelligence: With the help of various urban-infrastructure linked with the Internet of Things (IoT), the car’s AI system is capable of leveraging the cloud-hosted intelligence for autonomously doing various jobs for the driver. Starting from locating nearby gas stations or parking spots to embedded AI in cars, which can help navigation by utilizing data from the cloud.
iv. Superior in-car applications: AI-based infotainment systems in cars are enabling voice command based interaction for drivers and passengers to accept commands, search the internet, send and receive emails and also operate several smartphone apps without taking their hands off the steering wheel. Enabling Vehicle to Vehicle (V2V) or Vehicle to Infrastructure (V2I) technology to communicate with each other, several cars on any road would be able to maintain the minimum distance from one another for avoiding accidents, while the car leading into a kerb could also transmit its learning to the cars behind for smoother rides.
v. Smart insurance risk assessment: With AI systems built into the car, every minute detail of the driver’s nuances can be monitored while creating a risk profile which can be used for risk assessment of the driver and the vehicle by insurance companies. Smart automobile insurance powered by AI is finding its various applications in fleet management, logistics, and driver safety.

Benefits Of AI

• Data fusion: Data from diverse sources including in-vehicle sensors and data from external sources are combined to generate a surround view of the vehicle, which can be used by the autopilot functions to drive the vehicle safely.
• In-car driver assistance features: Adaptive cruise control systems are already in production. Along with lane-keep-assist and ability to pinpoint the vehicle with navigation data, it is possible to manoeuvre the vehicle with minimal driver interference. Ability to detect and prevent collisions and the ability to detect and understand traffic signs will enable the vehicle to do most of the driving by itself even in highly congested environments.
• Predictive maintenance: By analyzing the vehicle-generated data using cloud-based intelligent systems, predictive maintenance solutions that significantly increase vehicle uptime can be implemented.
• Self-diagnostic: AI systems can be used to provide self-help systems and assist in self-diagnosis. It’s possible to use mobile based applications to tag and provide data on real-time video feeds; making it possible to do assisted diagnosis on vehicles by pointing the mobile camera to appropriate sections of the vehicle.
• Usage-based insurance: By analyzing the driver behavior, it would be possible to provide usage-based insurance; with the majority of careful drivers paying less for insurance than an aggressive and more dangerous driver.
• AI systems for smart cities: Analyzing data from traffic cameras spread across the city, it’s possible to understand the area’s traffic situation for facilitating smart traffic management. The insights from analyzing parking areas can be used to reduce the time and effort required to park a vehicle.
AI is capable of creating a highly personal and customized driving experience by perceiving the car as something more than a tool for transportation. It assumes the role of a partner capable of anticipating the involvement level each person wants. On one day it serves as an extra set of eyes tracking pedestrians or cyclists in blind spots, while on another it checks the engine or brake oil status to fix an ‘oil change’ appointment.

Quest For AI

Renowned engineering service providers have recognized the disruptive and transformational role AI will play in the future automotive sector. Progressive providers are fortifying themselves for the future trends with strategic acquisitions and investments. However, predicting the future that would successfully leverage AI is tough for us. The technology has come of age. With the cost of storage reducing drastically, now it is possible to store and access large amounts of data. Computing capacity has incredibly increased to enable very high computing capacity in a mobile form factor (low power consumption, small size, and low thermal dissipation) to create the ability to do much more inside a vehicle. Even the ability to stream large amounts of data on the move has increased with the availability of 4G and 5G services. These developments have made it possible to use AI on automobiles to provide more and more features. Disruptive changes are on its way!
(Sherry C Kunjachan is the Strategic Initiatives Leader, Transportation, at QuEST Global. In this role, he focuses on growing scalable accounts and services for the Transportation vertical of the organisation. Views expressed are personal.)

Leave a Reply


This site uses Akismet to reduce spam. Learn how your comment data is processed.