How Artificial Intelligence is revolutionizing Transportation
Transportation has always been a crucial part of an economy: it has long been linked to the production, distribution, and consumption of goods and services. The oldest stock index still in use can be traced back to 1884. The Dow Jones Transportation Average was created when only two of the eleven companies in the index weren't railroad companies. Much has changed since then; not only has the Dow diversified, but the companies in it are relying more and more on technology to improve their businesses and stay competitive.
Many companies in the transportation sector have already started to adopt AI solutions, but they are barely scratching the surface of what is possible. The most well-known of these is the proliferation of Autonomous Vehicles, but more innovations can vastly benefit from AI over the next few years. Some companies in the transportation industry are also carving out part of their marketplace using an IP Moat, while others appear to be ignoring this crucial aspect.
Autonomous Vehicles
While there are many aspects of transportation being augmented by Artificial Intelligence, the most significant impact on individuals currently seen comes from Autonomous Vehicles. Tesla has become a household name, partially based on its Full Self-Driving feature, but it is not advanced compared to some of its rivals in the space.
There are six classes of automation for vehicles. They start at Level 0 with no automation whatsoever and increase to Level 5, which is so automated that it does not require a driver. Tesla and most other automakers are delivering Level 1 and 2 systems today. Audi produced the world's first production Level 3 with their 2019 A8L. The core difference between Levels 2 and 3 is the ability to use information about their environment to make informed decisions. Waymo, NAVYA, and Magna all have solutions that are Level 4. The improvement Level 4 adds being able to intervene when something goes wrong, or one of the systems in the vehicle fails. No Level 5 vehicles are currently available to the public. When they are available, they should drive anywhere safely and do not require human attention at all.
Safety is the primary way we will benefit from AI in autonomous vehicles. According to the National Highway Traffic Safety Administration (NHTSA), over 36,000 people were killed in 2019 in motor vehicle accidents, and 6.76 million total accidents were reported to the police[1]. Not all accidents can be prevented by Artificial Intelligence, but crashes related to excessive speed, alcohol impairment, drowsiness, and distracted driving accounted for almost two-thirds of 2019 motor vehicle deaths and would not be a factor with only fully Autonomous Vehicles on the road.
Most traffic stoppages are at least partially caused by accidents. Reducing them stops the biggest impediment to efficient flow. Safety also can lead to other benefits like efficiency and a reduction in energy usage. We have already seen incredible improvements in traffic routing related to solutions like Google Maps, which leverage AI in many ways, such as predicting traffic and determining how best to reroute vehicles when traffic changes. Traffic management will also improve data collection and decision-making to avoid traffic congestion. Energy used for transportation is a product of efficiency and will continue to decrease as efficiency rises. As energy increases in price, it will become more important as a factor to be considered.
Commuters that take public transportation can often take advantage of the time spent doing other tasks. Now commuters in their cars can gain back time in the same way that could improve productivity, shorten workdays, or relax with something entertaining.
Patent Landscape
Toyota has one of the strongest portfolios globally, and they target important domains with a clear focus. While slower to roll out autonomous solutions to the marketplace, Toyota leads everyone by a significant margin in granted patents using machine learning for self-driving vehicles. Tesla, the company many associates with self-driving vehicles, has taken the opposite approach. They have an order of magnitude less in the same domain, despite the advantage of crowdsourcing their learning data from all of its vehicles.
Looking at the families with active applications, we can see the group as a whole accelerating their filings significantly. Bosch and BMW both increased their investments enough to get them in the top ten, and Honda has more than double as many applications than granted patents in this technology, moving them slightly past Ford for the second spot.
Patent filings related to machine learning for autonomous vehicles (Source: Lumenci)
Disclaimer: Data for the past 2 years might not be accurate due to unpublished patent applications.
The real story here is the increase in filings over the last decade. Companies that understand the importance of Intellectual Property are rushing to secure what they can while other companies focus only on creating solutions and market share.
Intelligent Transportation Systems
A lesser-known area of advancement that will affect commuters and the rest of the Transportation Sector involves other portions of intelligent transportation systems (ITS). The main goal of such systems is to improve the safety and efficiency of moving people and goods. There has been steady progress towards this for decades. The U.S. Department of Transportation (USDOT) ITS Joint Program Office celebrated 30 years in 2021 and is focused on solving safety, traffic, and energy waste problems mainly by collecting and sharing all of the data available.
Vehicle-to-everything (V2X) will enable vehicles to communicate with any system that may affect, or be affected by, that vehicle. Imagine a system where all relevant data is available at all times. Some vehicles already have the required wireless technology built into them, but it needs widespread adoption to become functional.
Some solutions from researchers and companies are already using traffic data and AI to simulate traffic conditions and determine what actions to take. Smart Traffic Management Systems can be automated to optimize traffic flow by looking at real-time data and predicting issues. They will achieve that by controlling traffic lights, speed limits, on-ramp meters, highway message boards, or even reversible lanes. Eventually, they will have the ability to intervene by controlling vehicles to prevent accidents directly.
Some companies driving this technology forward are building Smart Traffic Systems, such as TrafficCom, Rapid Flow Technologies, Vivacity Labs, and even Google. Toyota has invested strongly here by creating the Toyota Mobility Foundation, which has goals that include "Improving access to connected transportation systems that are greener, safer and more inclusive by working with local partners and residents to design, deliver, and scale solutions for communities across the globe."
Others are using AI to make broader choices about transportation, like forecasting future demand for transportation infrastructure. NTT, Dillon Consulting, and IBM are examples of companies helping governments solve those problems.
Patent Landscape
Granted patents using machine learning for vehicle traffic control systems (Source : Lumenci)
Once again, the strength of Toyota’s portfolio shows well here. Two rideshare companies, Uber and Didi also have put themselves into good positions. The most interesting companies here in the Insurance industry are State Farm and Allstate. Both have sizable portfolios with patents in a wide variety of domains and investing in Intelligent Transportation Systems may benefit them in the future.
Current applications using machine learning for vehicle traffic control systems (Source: Lumenci)
Most of the same companies that got an early start are continuing their investments. Toyota still dominates, but Honda is moving strongly ahead of other competitors, and Hitachi, Denso, Panasonic, and NEC all pushed into the top ten.
Patent filings related to machine learning for vehicle traffic control systems (Source: Lumenci)
Disclaimer: Data for the past 2 years might not be accurate due to unpublished patent applications.
As shown in the graph, applications for patents related to using machine learning to predict or manage traffic have been trending upwards for the last decade. This is the case for most technologies that can benefit from machine learning as it has become more accessible and commonplace.
Supply Chain
Another huge impact will be felt when AI is pervasive throughout the supply chain. The AI will help streamline the movement of goods and materials to solve the current problems we face in the supply chain. Cargo trucks, trains, planes, and the shipping industry have all started benefiting from AI, but we will see a vast improvement over the next decade. AI will be seen at every level, from accurately predicting future demand to optimizing routes and automating warehouses and product inspection. All of this adds up to having the necessary products in the right place at the right time.
The shipping industry is a case study of how AI can benefit a big industry in transportation. It uses AI to automate much of the equipment used to move cargo. Problems specific to shipping, such as the impact of underwater fouling on performance, can be used to forecast performance. Similar to other transportation industries, navigation, safety, reduction of fuel consumption is being handled by AI. Examples across all these industries are already in use. Some of the companies pushing the boundaries here are Shone, Windward, Microsoft, Sea Machines Robotics, and Maana.
Large Autonomous transport trucks that can form efficient groups known as 'platoons' are being tested by multiple companies, such as TuSimple and Aurora. Practically every large airline use AI to predict how much fuel is needed for the flight, predictive maintenance, flight management, and a host of other processes.
In 2018, mining company Rio Tinto started using autonomous trains to move iron ore in Australia, developed with Hitachi Rail STS. Siemens and 4Tel also focus on increasing safety and efficiency in train systems with AI.
Patent Landscape
Patents related to the use of Machine Learning for improving parts of the supply chain cover too many domains, making it difficult to give a brief and informative look at the whole. Instead, take as an example TuSimple, which was founded in 2015. It began building a patent portfolio in 2017 and has filed more than 200 Patent Families since, with almost half of those relating to the use of machine learning in autonomous trucking. They are committed to competing in the market and aggressively protecting their place in it.
Contrast TuSimple with Aurora, a Self-Driving technology company founded in 2016. They understand the importance of the technology, specifically pointing it out on their website: "Machine learning is an essential component of our autonomy software, but we also know that machine learning alone is not enough to deliver safe self-driving." Despite that, their published patents are in the single digits, and only a couple of these teach machine learning methods.
Conclusion
Much of transportation as known today will completely transform in a few years. The pace of change is accelerating, and everyone stands to benefit from them. Many companies are already well-positioned technically to compete in these industries, though some will struggle because they lack sufficient dedication to protecting their IP. The organizations that do both will dominate in the marketplace.
*Disclaimer: This report is based on information that is publicly available and is considered to be reliable. However, Lumenci cannot be held responsible for the accuracy or reliability of this data.
*Disclaimer: This report is based on information that is publicly available and is considered to be reliable. However, Lumenci cannot be held responsible for the accuracy or reliability of this data.
Editorial Team at Lumenci
Through Lumenci blogs and reports, we share important highlights from the latest technological advancements and provide an in-depth understanding of their Intellectual Property (IP). Our goal is to showcase the significance of IP in the ever-evolving world of technology.