Original article was published by Angelo Dalli on Artificial Intelligence on Medium
How AI Impacts Transportation
How will the transportation systems of the future look like? Are we going to have driverless cars everywhere? Are we going to have trucks delivering things to our homes directly? Are we going to have little drones dropping packages next to our houses? Or will the flying cars and drones of the future be too noisy for us to allow them to come too close to our houses — because let’s face it, no one wants jet engine noise or propellers constantly above their heads! Are there going to be things like no-fly zones over houses and over streets? And maybe limited times when the drones can go and supply things to you, and when people can use personal flying vehicles — because you don’t want your neighbours to be coming and landing their big flying car in the middle of the night and waking you up! When people imagine how future transport systems and AI will look like, a lot imagine advanced driverless cars and sophisticated ways of getting from point A to point B quickly, safely and cheaply.
The best driverless systems currently available can be more accurately described as being driver assist systems rather than completely autonomous driverless. Driver assistance features and partially autonomous features are going to become more common in all cars, and this will read in a reduction of accidents and possibly faster transit times due to decreased congestion and better utilisation of the transportation infrastructure. AI can also help us achieve a safer transport system in view of pandemics and disease, by maintaining cleanliness while avoiding unnecessary contact without needing restrictive lockdowns by ensuring safe compliance with the relevant rules appropriate to the situation at the time.
Public transport systems already have some automated systems in use, especially in fixed-track systems like metro trains. I think that public transport will become increasingly automated with automated buses, vans, trains and ships. When it comes to planes, for example, we might start seeing planes with a single human pilot, and an AI pilot that will effectively be the co-pilot, and a third human backup who will be on the ground, controlling the plane remotely if need be. The third human pilot can be shared across multiple flights and hopefully will not need to spring into action frequently, as the AI can take over during pilot breaks and in case of emergencies requiring quick on-site reaction. When it comes to shipping, having an AI system that can dock a ship or that can pilot it to its remote destination using less staff than is needed now will make a big difference. Making sure that there will also be cargo transport that is completely automated will have a profound impact on our society. Things like hyperloops and other innovations will also make sure that high speed can be achieved safely, especially since automated systems can driver at higher speeds and avoid the risk of collisions. Future systems may have segregated traffic, with automated traffic having its own high-speed lanes. Even though many people seem to think that AI is on the cusp of being everywhere in transportation there are still several issues to be resolved for the ideal transportation system of the future to be achieved. One of the current issues holding back progress is the sheer amount of data that AI needs to learn. For example, in certain situations such as when driving in deep snow, and when encountering new and previously unseen situations, AI systems currently cannot adapt fast enough like a human driver would.
How will we interact with such driverless cars and driverless automated systems? How will society interact with transport AI in a collaborative manner? If you are stopped at an intersection or you are just crossing the street and see a driverless car coming along, how will you know for sure that the car is going to stop for you? How will the car indicate to you that you can pass, and that it is safe to do so? These kind of issues need to be resolved and having AI that can explain itself and that can interact effectively with people and other autonomous systems — AI that is in general trustworthy — is going to be a key part in achieving the seamless and hassle free future that we all deserve.
Apart from driverless cars, AI can help in other ways in transportation. AI can help us keep fuel prices affordable, help us carry out better research into alternative fuel sources, for example — electric cars and hydrogen fuel cells may be the main sources of transportation fuel in the future. AI can also help us design better and more efficient engines and transportation vehicles using generative design and other related techniques. The transport infrastructure can be maintained in an optimal manner by a maintenance fleet of drones and robots that will constantly monitor the infrastructure and its usage status, quickly reporting where there are potholes that need to be fixed, signs that need to be replaced, sites that need to be maintained and similar kind of tasks that can really benefit from predictive maintenance and pro-active action. Using AI to do this kind of monitoring and repair can significantly save on the cost of maintenance, while ensuring a better and safer experience for everyone. The same maintenance fleet can also detect accidents and ensure that transport users needing assistance receive it in a timely and effective manner.
AI is a key enabler of the intelligent transportation system of the future, a transportation system that is quicker, faster and makes best use of our limited resources, helping us keep the environmental impact as low as possible while enabling us to enjoy more time productively or simply having fun connecting with other people while travelling from one place to another.