For years, the future has been envisioned with flying cars. While one day this may become a reality, it is not the case today. However, truly autonomous, self-driving vehicles certainly are a present possibility. Given the manner in which AI is beginning to dominate the world, it may well soon be a reality. However, although the technology exists – and we have all seen demos and heard tales of self-driving cars – it is difficult to picture our roadways filled with them.
As a society, the concept of driverless vehicles is still, essentially, a novel concept. Not only are driverless cars scare currently, but the regulatory and legal ramifications of their existence are still being worked out.
Who is legally at fault if a driverless vehicle is involved in an accident or causes damage to persons or property? Is it the owner, who wasn’t in control of the vehicle? Or is it the manufacturer’s fault? Or is the service provider or vendor who controls the vehicle’s autonomy system and database to blame? On these fronts, there is still a long way to go, at least until we can decide upon the finer details.
Driverless vehicle technology is forecast to contribute $7 trillion to the world economy over the coming decades according to Intel. It will also save thousands – perhaps millions – of lives. Public sentiment remains ambivalent, however, despite the increase in uptake. For example, the type of hardware used to make these vehicles possible and operational is not what may be expected. Most people would imagine robot-like controls which are connected to a larger track-based or chassis system in which all vehicles run on a roadway which is reminiscent of a theme park.
In actuality, data will power the future’s driverless vehicles. Digital information or data collected by an array of monitoring systems and sensors which is then passed through a big data platform connected to machine learning or AI will be used. It seems far-fetched, particularly considering that digital content or “data” is not tangible in nature. How is it possible for an invisible, impalpable element to have control over something as advanced as driverless vehicles, and by proxy people's lives?
The Relationship Between Self-Driving Vehicles and Data
Put bluntly, without their connected data, driverless vehicles would not exist. It just wouldn’t be possible to make them. This is because all of the situational and contextual information from the real-world must be collected, processed and deployed in order for automated systems to carry out their work. Driverless vehicles must remain constantly vigilant about other road users, nearby objects and pedestrians, and the direction it is going and how that relates to its present position.
The required information is collected by numerous devices and sensors, which then transfer it via an open connection to a main system or hub and then receive it once again after it has been collated and processed. Digital data into transformed into usable insights by that central system, which is frequently powered by AI or machine learning algorithms.
Crucially, this process must be completed at such a fast speed that the control unit of the vehicle has enough time to send and receive the related information prior to making a split-second decision. It needs to be able to process whether an object rolling in the road in front of it – at a distance of less than a mile – is a ball or a child. Additionally, it must know whether it has enough time to stop, should swerve out of the way, or should maintain its course.
This means that, even in simple situations, the system is responsible for a lot of lives. It is responsible for keeping safe the passengers inside the vehicle, as well as anyone in the surrounding area – like the child who lost their ball.
Even if it can be difficult to comprehend, data drives it all.
How Self-Driving Vehicles Utilize Data
Regarding general importance, the data which is processed and used by driverless vehicles is potentially either vital or inconsequential. For example, being aware of the situational and environmental surroundings is potentially the difference between reliable operation and a major accident.
However, data about the car's passengers can also make a difference, although not in the same way. Information about passengers could be used to alter the vehicle's temperature control settings or entertainment devices, as well as dynamic accessories such as an electronic window tint system.
Situational information is pivotal to the car's main operations and journey. Data collected about roadways, structures, or nearby landmarks are able to be used to discern the car's current location. In combination with modern GPS information, this process would enable the car to precisely pinpoint its geographical position.
In this industry, geographic information systems are used in a number of ways. They are able to aid in the planning of routes, the identification of shortcuts, and how best to avoid traffic. They are also able to sync information to remote servers, which could relate it to other vehicles nearby on the same roadway.
Accidents occurring just minutes ahead of the car could be relayed to a nearby vehicle, thus enabling it to slow or stop as required. Additionally, the vehicle would be able to see, gauge, and measure surroundings constantly, courtesy of a sonar-like sensor.
Vehicle systems are able to see their present location thanks to the external body, speed, hardware and performance information. When dealing with environmental issues, this could become even more important. As an example, avoiding flooded roadways requires the vehicle to continuously measure the water levels touching the chassis in order to identify whether or not the water is shallow enough to be crossed.
Data is also being utilized in order to prepare both our roadways and the vehicles on them for the arrival of driverless cars.
Operational Testing and Deployment Data
Prior to the release of these vehicles, society must be prepared and manufacturers must be certain that they are to work as intended. A great deal of testing is required for that level of assurance, both on real roads and in real cities.
In order to aid the optimization of hardware and systems used, all the data being collected is also being fed into deployment and testing. For example, if sensors have a blind spot, they must be augmented with additional devices or retooled in order to fix vulnerabilities. The only way of ascertaining whether or not such a device is measuring accurately on real roads is to test it out on them.
As much as two Petabytes – or two million Gigabytes - of data is created and processed by Google’s self-driving vehicles every year. That is a great deal of data. Some of this data is local and technical – such as what the car does in varying circumstances – while community or external data, as well as personal data concerning the car’s passengers, also exists.
However, the most important thing to remember is that these systems and vehicles don’t stop collecting, processing or using data until they are parked and turned off. Even at that point, however, data continues to be used in order to increase the power, accuracy, and safety of the systems.
If you have enough data, it is possible to begin deploying predictive systems which make correct decisions which are based on fact. If a ball rolls out onto the street, it is quite likely a child or pedestrian will soon follow it. If possible, come to a complete stop. If not, then swerve. It was best put by Intel CEO Brian Krzanich, "Data is the new oil." Vehicles of the future will be fuelled by it, as well as being made more aware, efficient, and smarter than they have ever previously been.
Enter Privacy and Security
With this huge scale of data collection and its free flow between systems, open connections, and deployment, there are concerns in relation to security and privacy.
For one, how are we able to prevent hackers from accessing the data control systems and wreaking havoc? What happens if you are on a highway and your vehicle gets shut-down by an external party? Some severe or fatal accidents could result from this issues, which could have repercussions both for the passengers of the vehicle which was shut-down and other passengers on the same road.
Another concern is privacy. If the vehicle is collecting and reporting information about passengers and drivers at all times, what does it see? Roadway signage and billboards could, for instance, be directly influenced by this data. Similarly to Amazon and other retailers’ use of your browsing history to suggest products, a dynamic system could pick up good or products you are discussing and then display them on signs nearby to boost advertising success rates.
As important as these factors are, it will, unfortunately, take time to figure out the kinks of the technology. It is also necessary for us to be vigilant concerning how our privacy and personal data are handled by the companies which build these vehicles. Consumers must be protected, even from necessary and helpful systems, such as those powering driverless vehicles. Research, awareness, proper regulation, and time are some of the necessary requirements – however, it won’t happen overnight.
This information has been sourced, reviewed and adapted from materials provided by Kolabtree, originally written by Nathan Sykes. Nathan Sykes is a freelance tech writer from Pittsburgh, PA. He enjoys writing about the latest news and trends in AI, big data, cloud computing and other emerging technologies.
For more information on this source, please visit Kolabtree.