Our Future with Autonomous Vehicles

From New Media Business Blog

Jump to: navigation, search

Contents

Introduction to Autonomous Vehicles

What are Autonomous Vehicles?

CMCV.
Girl in an Autonomous Vehicle [1]

An autonomous vehicle is a vehicle that is capable of navigating without human conduction by sensing its environment. [2] These vehicles have become a prominent reality and have begun to pave the way for future systems in which human input is no longer needed and computers are able to take over the art of driving. Specifically, driverless cars are a type of autonomous vehicle that have rapidly began to dominate the autonomous movement. [2] Companies are particularly focused on perfecting their version of an autonomous car; spending billions of dollars and countless resources in the race to create a fully autonomous, commercialized car.

Autonomous Levels

The word autonomous, however, holds a specific definition. The US Department of Transportation's National Highway Traffic Safety Administration (NHTSA) defined five different levels of autonomous driving in 2013. [3] In October 2016, the NHTSA updated their policy to reflect that they have officially adopted the levels of autonomy. [3]

Level 0 - No Automation

Level 0 is the most basic level. The driver controls it all: the steering, braking, throttle, power, etc. [3] It is how almost everyone in 2018 currently drives.

Level 1 - Drive Assistance

In Level 1, most functions are still controlled by the driver but one specific function can be done automatically such as steering or accelerating. [3] For example, a car that brakes a little extra when it gets too close to another car on the highway.

Level 2 - Partial Automation

In Level 2, at least one driver assistance system is automated. [3] It means the driver is disengaged from physically operating the vehicle by having their hands off the steering wheel and foot off the pedal at the same time, for example. [3] However, the driver must still be ready to take control of the vehicle and the major aspects of driving are still handled manually. [4] Most current automakers are developing vehicles at this level, where the vehicle can assist with steering or acceleration functions and allow the driver to disengage from some of their tasks.

Level 3 - Conditional Automation

Starting at Level 3, the vehicle itself controls all monitoring of the environment (using sensors like LiDAR). [4] In this level, a driver is still necessary but are able to completely shift “safety critical functions” to the vehicle, under certain traffic or environmental conditions. [3] It means that the driver is still present and will intervene if necessary, but is not required to monitor the situation in the same way it does for the previous levels. In fact, many current Level 3 vehicles require no human attention to the road at speeds under 37 miles per hour. [4] Interestingly, many large car manufacturing companies such as Audi have announced the release of their own Level 3 autonomous cars to launch in 2018. [4] An autonomous vehicle expert at Ford noted that they plan to take the company straight to Level 4, saying “We’re not going to ask the driver to instantaneously intervene — that’s not a fair proposition.” [4]

Level 4 - High Automation

Level 4 is what is meant when people speak of “fully autonomous.” The vehicle is capable of steering, braking, accelerating, monitoring the vehicle and roadway as well as responding to events, determining when to change lanes, turn, and use signals. [3] They are "designed to perform all safety-critical driving functions and monitor roadway conditions for an entire trip." [3] However, it's important to note that this is limited to the operational design domain (ODD) of the vehicle—meaning it does not cover every driving scenario. [3]

Level 5 - Complete Automation

Level 5 refers to a fully-autonomous system that expects the vehicle's performance to equal that of a human driver in every driving scenario. [3] The vehicle requires no absolutely no human attention. This includes extreme environments like dirt roads that are unlikely to be navigated by driverless vehicles in the near future. [3] There is no need for pedals, brakes, or a steering wheel, as the autonomous vehicle system controls all critical tasks, monitoring of the environment and the identification of unique driving conditions like traffic jams. [3]

It is important to know that these levels of autonomy describe the system, and not the vehicle itself. They are general guidelines for how technologically advanced vehicles are and are important in terms of what consumers need to know.

How Do Autonomous Vehicles Work?

How Do Driverless Cars Work? [5]
CMCV.
Components of Autonomous Vehicles [1]

For fully automated driving to prevail, driverless vehicles need to exceed the capabilities of human drivers. Different types of sensors enable the vehicles to make sense of their environment, in a more foresighted way than a human ever could.

Image Sensors

Several cameras imitate human eyesight and generate images of the vehicle's surroundings. [2] Cameras can detect colours and fonts which is extremely important to be able to interpret traffic signs, traffic lights, and lane markings.[2] They also act as a backup system if other sensors on the vehicle fail. [2] In order for the three-dimensional images to be realistic, four to six cameras are required. [3] Some current challenges with image sensors are that they cannot see far enough to enable more anticipatory driving. [2] In addition, certain weather limitations such as fog, rain, or low sun seem to affect the cameras interpretive ability.[2] These limitations increase the risk of failures and need to be overcome.

Radar Sensors

Radar stands for “Radio Detection and Ranging”, meaning these sensors detect and localize objects using radio waves. [3] If the waves hit an obstacle, the waves are reflected, revealing how far away an object is and how fast it is approaching. [3] Short and long-range radars are deployed all around the car and they are able to track the speed of other vehicles in real time. [3] They also contribute to safety by increasing the redundancy of autonomous driving systems. [3] Some other radar sensor applications include: [3]

  • Blind Spot Detection (Blind Spot Monitoring)
  • The lane and the lane-change assistant
  • Rear end radar for collision warning or collision avoidance
  • Park Assist
  • Cross-traffic monitoring
  • Brake Assist
  • Emergency braking
  • Automatic distance control

    Lidar Sensors

    LIDAR is an acronym for Light Detection And Ranging and is a laser-based system. [3] These non-visible laser beams require a highly sensitive receiver and are primarily used to measure distances to both stationary and moving objects. [3] The low intensity, non-harmful beam visualizes objects and measures ranges creating a 3D image of the cars environment. [3] Combined with the data of camera sensors, LIDAR sensors also allow obstacles to be accurately identified, for example determining if whether a vehicle or a pedestrian is in front of the car. [3] As mentioned, the challenge for driver assistance systems is ensuring that the system will successfully function under all environmental conditions such as temperature, solar radiation, darkness, rain, snow, etc., in addition to recognizing objects up to 300 yards away. [3] Subsequently, another challenge is to produce at the lowest possible cost and smallest dimensions for large-scale production. Unfortunately, at costs of up to tens of thousands of dollars, these mechanical systems are currently not suitable for large-scale deployment in the automotive sector. [3]

    Cloud

    Another newer feature of autonomous vehicles is the Cloud. The cloud offers highly accurate, real-time maps that vehicles draw on, serving as a dynamic electronic horizon. [2] The data is constantly updated by the collective intelligence of driverless vehicles, for example reporting closed lanes or defective traffic lights. [2] Therefore it can also be considered a type of sensor - providing the vehicle with an image of its surroundings.

    History of Autonomous Vehicles

    The history of autonomous vehicles surprisingly began sooner that one may have thought. There are several historical events which shaped the way autonomous vehicles have become a prominent reality in everyday lives. Some of the most notable vehicles include da Vinci's Self-Propelled Cart, the Linrrican Wonder, the Mercedes Robot Van, and Google's Waymo project.

    da Vinci’s Self-Propelled Cart (1478)

    CMCV.
    Da Vinci's Self-Propelled Cart (1478) [4]

    Centuries before the invention of the automobile, Leonardo da Vinci designed a cart that could move on its own. [5] The springs put under high tension provided power to the cart while steering could be set in advance so the cart could move along a predetermined path. [5] A distant precursor to the car, the device is sometimes considered the world’s first robot. [5]

    Linrrican Wonder (1925)

    One of the first widely known attempts for driverless cars was the Linrrican Wonder. It was produced in 1925 by Houdina Radio Control, a radio equipment firm. [6] The Linrrican Wonder traveled through a traffic jam in New York City and was controlled by a transmitting antenna. [6] The car behind this “phantom auto” sent signals to the antenna, and the signals worked on small electric motors. [6] The car could start its engine, shift gears, and sound its own horn. [6]

    Mercedes Robot Van (1980)

    In the 1980s to 1990s, Mercedes-Benz began experimenting with driverless vehicles. [7] Daimler-Benz, in cooperation with several European car manufacturers, launched the Eureka PROMETHEUS (Program for European Traffic with Highest Efficiency and Unprecedented Safety) Pan-European project. [7]

    CMCV.
    Mercedes Robot Van (1980) [7]


    It was then the largest research and development program in history for autonomous driving vehicles and related technologies with a total cost estimate of 749 million euros in current money. [7] Mercedes-Benz made a number of driverless prototypes with the main idea behind the project being not to remove the driver from the equation, but to offer an accident-free driving experience. In fact, they built the world's first robot van which on an open road achieved a speed of 39mph. [7]

    Waymo (2009)

    In 2009, Google began developing its self-driving cars, but did so privately, avoiding public announcement of the program until a later time. [8] Google started its project with six Toyota Priuses and an Audi TT that drove through the streets of Mountain View, California. [8] By 2010, Google's cars had logged more than 140,000 miles on the road. [8] The company wrote in a blog post at the time that they believed Waymo had the power to cut the number of traffic deaths each year by half. [8] The company also added several high-profile engineers to its team: Carnegie Mellon University professor Chris Urmson, who led a team that won the 2007 Urban Challenge; Mike Montemerlo, whose Stanford team won the 2005 Grand Challenge; and Anthony Levandowski, who built the first autonomous motorcycle. [8] In May 2014, Google built its own car and showed off a prototype of it at the Code Conference. [8] There were no brakes, no steering wheel, and no gas pedal — only a button to turn it on. [8] The company capped the vehicles' speed at 25 mph and planned to make 100 to 200 versions of the car. [8]

    Where Autonomous Vehicles Stand Today

    As we move into the 21st century, companies are beginning to invest more and more into the driverless car industry. Technological advances in the industry are growing at a rapid pace. It is predicted that nearly 10 million cars with self-driving features will be on the road by 2020. [9] And just a couple of years after 2020, its predicted that we’ll see fully autonomous vehicles that can drive on roads and handle a range of scenarios with little or no interaction from the driver. [9]

    Naturally, there are a few issues to overcome before autonomous vehicles are your everyday ride. Along with technical hurdles, there are legislative barriers, insurance and liability questions, and even cybersecurity concerns. At the moment, only a few American states — including California, Florida, Michigan, Phoenix and Nevada — allow self-driving cars to be tested.[9]

    Autonomous Vehicle Concerns

    Safety

    The principal goal of the self-driving movement is to make roads safer by eliminating human error.[10] It is predicted that autonomous vehicles could prevent anywhere from 75% to 90% of the auto accidents we suffer each year. [10] But taking the reins from the person in the vehicle requires a lot of programming for quite a few dilemmas.

    Harsh Weather

    Currently one of the main issues with safety that driverless car producers are facing is harsh weather. [10] Right now, most self-driving cars are stumped by harsh weather.[10] Heavy rainfall interferes with sensors, and snow on the ground makes it hard for vehicles to read lines on the road.[10] There are also little safety things like misreading harmless puddles of water as potholes and slowing down for no reason.[10]

    Laws and Regulations

    While there are many areas, as mentioned previously, that welcome the possible economic benefit of lax regulations, many other areas are responding with healthy skepticism, expecting companies to comply with developing laws.


    In Canada, a new report from the Senate Committee on Transportation and Communication says this country is not prepared for the new technology and measures must be taken to ensure safety. [11] The Senate report urges regulations to ensure the safety of driverless vehicles, especially during Canadian winters when snow covers road markings and changes the handling properties of the vehicle. [11] Currently, the Senate report has 16 recommendations, including a national strategy for autonomous vehicles, safety guidelines, and cybersecurity measures to ensure the country is ready for the next era of transportation. [11]

    CMCV.
    The Tesla Model X P100D with Ludicrous Mode starts at $140,000 [12]

    Price

    Currently, we see driverless cars as a luxury item. However, as mentioned before, the goal of the driverless-car movement is to eliminate human drivers completely. The biggest benefits of driverless vehicles can be attained only when human error is removed from our roads. Monopolizing driverless functionality would hinder the tech's chance to cut down on accidents and auto-related deaths, because streets and highways would still be dominated by people.

    Liability and the Coding of Autonomous Vehicles

    Another hurdle for self-driving cars will be liability. Currently, when traffic accidents occur, the driver is considered liable, not the vehicle. However, with driverless cars becoming a prominent reality, consumers and manufacturers may begin to wonder who will be liable in a driverless car accident. This has been an ongoing debate with many examples of real-life cases. In 2017, an Uber operated self-driving car killed a pedestrian. [13] It appears to be the first time an automobile driven by a computer has killed a human being by force of impact with the car driving at 38 miles per hour. [13] As the investigation opened, police indicated that the pedestrian may have been at fault but with minimal evidence. The future of autonomous car liability is still unclear and may vary case by case. [13]

    Conscious Decisions

    As autonomous vehicles gain popularity in our society, questions around ethics come up. If a vehicle crashed in the past, it was a reaction by the driver and not a conscious decision.[14] It was instinct, and society was sympathetic because humans aren’t perfect, and we make mistakes. The use of AI and driverless cars complicates this. Decisions about the results of accidents are made with the programming of the car's driverless system months or even years in advance.[14] So the question is, who should make these decisions? Should it be the programmer? Should it be the company? Should it be the driver? Or perhaps the government?  


    Below is a list of pros and cons for each of the groups:


    Group Pros Cons
    Programmer Programmers are directly involved in the creation of the AI and may want a say in the creation of its decisions. Furthermore, they may not feel comfortable implementing something they don’t believe in. This would result in no standard practice for businesses being developed and programmers would feel pressure by the company to compromise their beliefs to increase profits.
    Company Customers will be happy because the cars will protect them over others. Will choose what customers want – protection of driver over common good.
    Driver May increase adoption of the technology if drivers are given a choice. Most drivers will choose their own protection over the common good.[15]
    Government Would result in a standard practice that could be enforced. Will lean towards the common good. Will take a long time to decide and debate in order to come up with a plan.

    Human Error

    When thinking about driverless vehicle ethics, it is important to remember the advantages of AI technology over human drivers. In the picture below, we can see that a driverless vehicle can stop in the approximate time a human takes to even notice there was a problem, thus exemplifying why this technology is a hot topic and many people want it implemented as soon as possible.

    CMCV.
    Stopping Distance AI vs. Human Drivers[16]


    The Social Dilemma of Driverless Cars [15]

    Worldwide, 1.2 million people die per year in traffic accidents and the main source is Human Error.[1] The longer we wait to get this technology on the road, the more people will die. This is a larger ethical issue about the concept of driverless cars than their decisions. If say, autonomous vehicles could prevent 90% of accidents is it moral to keep the tech off the road because it is not 100% safe?[1] Or is it moral to prevent these 90% of accidents happening by implementing the tech earlier even though it’s not fully functioning in all scenarios?[1] If we decide to implement the technology before it is 100% safe, at what level do we pick a different decision? 85%? 80%? 70%?


    This Ted Talk video to the right explains this concept and outlines how implementing driverless cars is in fact a “social dilemma”, where society must decide on and enforce trade-offs they are comfortable with.[1]

    The Trolley Problem

    The Trolley Problem [2]

    In 1967, Philippa Afoot devised what is known as the “trolley problem”.[1] The original thought experiment included a run-away trolley, headed towards 5 workers.[1] A bystander, who is the one doing the thought experiment, can pull a lever to divert the train onto a different track.[1] The problem is however, there is another worker on that track as well.[1] Should the bystander pull the lever and save more lives? Or should they “do no harm” by not intervening, which results in five deaths?

    Saving the greatest number of people is consistent of the philosophical principle of Utilitarianism “which argues that the morally correct decision is one that maximizes well-being for the greatest number of people.[1] In tests and virtual reality simulations, about 90% of participants chose this option.[1] It is interesting to note, however, that as the trolley problem has been expanded to different scenarios, people’s choices are not always logical.[1] When asked if they would push a large man in front of the trolley to stop it, only 10% of the participants chose that option.[1] Under Utilitarianism, the choice should be the same, but it’s complicated by the fact “that pushing someone to their death feels more personal” than pulling a lever, and that creates emotions which cause participants to avoid the option.[1]

    In the future, driverless cars will have to make decisions like the trolley experiment. For example, they will be required to make the decision of causing a small accident to avoid a large one.[1]

    If you would like to judge trolley problem simulations and compare your preferences to others, check out the Moral Machine.

    Hacking

    Remotely Hacking a Jeep [2]

    Due to the impending rollout of autonomous vehicles on a national scale, the companies that manufacture them will need to be wary of the unseen dangers that could prove to be fatal. In the 1990s, there were no reported hacks when the dot-com startups first released their e-commerce platforms, however, following the first series of hacks to these sites, Microsoft started taking security seriously. [1] Today, their operating system, Windows, is one of the world’s most secure operating systems with Microsoft now spending more than a billion dollars annually on cybersecurity. [1] Yet despite these immense efforts, hackers are constantly finding ways to disrupt Windows operating systems, browsers and applications. Like the pc, autonomous vehicle hackers would have malicious intent behind their attacks. But unlike the pc, the dangers go far beyond just stolen personal and financial information.

    It would be naive for autonomous car companies to believe that their cars will not be susceptible to cyber criminals. In fact the magazine, Wired, ran a field test story just to show that even cars today can be hacked remotely. [2] Security researchers, Charlie Miller and Chris Valesek expose the security vulnerabilities in automobiles by remotely hacking and taking control of the cars' various controls from the brakes to changing stereo music. [2] After the story ran, Chrysler recalled 1.4 million of their vehicles, suddenly aware of the assortment of dangers within their vehicles. [3]

    CMCV.
    Security Vulnerabilities within Present Day Vehicles [4]

    Within the internal systems of autonomous cars, it is relatively easy to see just how vulnerable the car is. Malicious hackers could gain access to everything from the critical sensors that detect other cars and dangerous hazards to the steering mechanisms controlling the car. Case in point, imagine a scenario where a hacker takes over a vehicle and forces the passenger(s) to perform a crime like robbing a bank or murder; otherwise, they would crash the car. In an escalated scenario, the hacker could even run over the victim if they managed to escape once out of the vehicle. Other scenarios include stealing financial or identity information stored in the vehicle databases. [5] Keeping the notorious 2013 Yahoo data breach in mind (3 billion Yahoo accounts were hacked), expert hackers could potentially launch a similar mass attack on millions of autonomous vehicles. [6]The results would borderline on a cybersecurity disaster.

    There are a variety of different attacks in the arsenal of savvy hackers. They could attack the GPS navigation on the vehicle, possibly rerouting the car into dangerous traffic or destinations. Through the onboard internet browser, they could send a virus through downloaded applications or sites to gain access carrying financial information such as credit card numbers. [4] Examples would include bank apps, online shopping, or simply any systems requiring an internet connection. [4]

    Additionally, hackers could take control of vital systems such as the Can-bus. The Can-bus is the computer’s own central networking system that interconnects virtually all ECUs (Electric Control Units) and sensors. [7] ECUs control things such as the audio and the engine control unit. [7] This gives hackers access to intercept the communication between the two and rewrite the signals into doing different actions via the software.

    In terms of defending hackers, some development has begun in research facilities worldwide. For example, at MCity, the University of Michigan’s dedicated autonomous car research and testing facility, they have a developed a defense matrix for testing on autonomous vehicles. [8] The matrix identifies threats and subsequent solutions based on the motivation, variability of the threat, target of threat (ex. Navigation system), and other measures. [8] Following calculations, the score is weighed against a board of likelihood of occurrence and then the result will reflect possible solutions for the CPU to commence based on the assumption of attack. [8] For example, the picture below shows the result of a foreign attack, displaying three different areas of attack and the area of highest of impact and attack potential. The “7” below shows that the spoofing attack (foreign communication acting as a known communicator to the receiver) [9] is of highest impact and highest potential. Given this information, the CPU would then prioritize and commence defensive actions. [8]

    CMCV.
    MCity's Prototype Hacking Defense Matrix[8]

    The goal of ultimate safety for passengers will now go beyond the vehicle’s structure and driver aids (back-up camera, blind-spot viewer, auto-emergency braking). It will now require a system that protects the passenger physically and the very system that runs the car from hostile cyber takeovers.

    Major Technology Companies and Players

    Tesla

    CMCV.
    Tesla Model S [10]
    CMCV.
    Tesla Roadster [11]

    Tesla is an independent car maker based out of Palo Alto California that was established in 2003. Named after the famous electrical engineer and physicist Nikola Tesla, the company specializes in making electric vehicles, lithium ion battery energy storage and solar panel manufacturing for households and businesses.[12] [13] The company's goal when it was founded was to prove the world that consumers did not have to compromise in order to drive electric vehicles, but instead they could be better than gasoline vehicles. [14]

    Currently, the company sells the vehicles Tesla Model S, Model X, Model 3, as well as the energy products Powerwall, Solar roof and other commercial products that focus on clean energy and energy efficiency.[12] All these products align with the company’s goal when founded, as well as the lack of fossil fuel use and wasteful emissions. Tesla’s vehicles all focus on combining safety, efficiency, performance and comfort.[12]

    To expand the company’s offerings they have the Tesla Roadster, their attempt at a supercar. The roadster features a top speed of over 400 km/h along with three engines, and is scheduled for a 2020 release.[15] In addition to their car lineup, they will also be offering the Tesla Semi, a fully electric trailer truck which would cater to the operations and delivery area of business.[16] It features four engines, low cost of ownership compared to diesel trucks, and multiple safety features. [16]

    The company consistently innovates with different products, making giant strides for the electric vehicle industry in the past 10 years. From its initial public offering on NASDAQ in 2010, to its increasing sales over the years, Tesla is here to stay.[12] Most recently in June 2018, Tesla reached the milestone of producing over 5000 Model 3 units, along with 2000 Model X and S, for a combined 7000 vehicles in a single week. [17] If the company were able to maintain this level of production, they would total over 300,000 electric vehicles per year, which is still unheard of today compared to the gasoline vehicle industry. [17]

    Autopilot

    Enhance your commute with Tesla Autopilot [18]

    Tesla vehicles feature full self driving hardware which enables the vehicles self driving capabilities at a safety level that is “substantially greater than that of a human driver”.[1] The vehicles have a total of eight cameras that provide a full 360 degrees of visibility for up to 250 metres.[1] They also feature 12 ultrasonic sensors that allow for detection of hard and soft objects, such as walls and humans respectively. [1] Additionally the car has a radar that provides data on the cars surroundings even through heavy rain, fog, dust, and even beyond the car ahead. [1]

    The onboard computer has enhanced software for vision, sonar and radar processing. [1] This allows the vehicle to have a view of the world that drivers cannot see, seeing everywhere simultaneously and even further than human senses can. [1] Autopilot features include matching speed to traffic, lane control, switching lanes without driver input, switch between freeways, self parking, and even being summoned to and from a garage, making it even easier for users to park their cars. [1] Tesla states that drivers are responsible for being active and ready to take action when using autopilot, which ensures drivers are aware and take responsibility as soon as possible. [1]

    The full self driving capability of the vehicle provides twice the safety of the average human driver. [1]It will allow drivers to do long and short distance trips with no action required, and for cars with superchargers with automatic charge connection, the car will charge itself. Additionally, the vehicle will find the optimal route to your destination, drive you to it, and find parking automatically. Once you are ready to leave, you can summon the vehicle through the Tesla smartphone app, and the car will come to you. Self driving capabilities are dependent on approval, and they vary between countries and the regulations of each country. [1]

    Google's Waymo

    CMCV.
    Google's 'Waymo' - Self Driving Car [2]

    Google’s self-driving car project started in 2009 and has now become one of the biggest competitors in the autonomous vehicle industry. [3] The company started on one premise: to build an autonomous vehicle by 2020. [3] Now in 2018, Waymo’s fleet has self-driven more than 7 million miles on mostly city streets, in addition to the 2.7 billion miles they drove in simulation in 2017 alone. [3] While Waymo has been successful in many ways, they have also faced some failures. Google has failed to commercialize its cars so far, meaning they have fallen behind competitors like Uber and Tesla. Between 2009 and 2015, Google spent over $1.1 billion on the project. [4] This is comparable with the $1 billion acquisition of Cruise Automation by General Motors in March 2016, [5] as well as the same investment by Ford in a joint venture with Argo AI in February 2017, [6] or the $680 million investment for Otto shelled out by Uber in August 2016. [7] Today, Waymo is worth $70 billion. [8]

    Uber

    CMCV.
    Uber Driverless Volvo XC90 [9]

    Uber had plans of creating a mass fleet of automatic cars which started with the purchase of 20 Ford Fusions in 2015. [9] The cars were fully equipped with a lidar unit, cameras, lasers, GPS and radars on the roof of the car.[9] The company ran multiple tests throughout the United States, after running many tests with the Ford vehicles, Uber went ahead and chose the Volvo XC90 as their automated vehicle of choice, which led to the purchase of 100 of them. [9]

    Unfortunately, in March of 2018, an Uber self-driving car struck a pedestrian, who later on died from their injuries. [9] This accident led to the immediate suspension of Uber’s public testing in the United States. [9] Ever since this accident, the future of the uber self driving program has been in the limbo.[9] All tests are suspended with no information on when they will resume. [9] Uber had made a large purchase of 24,000 vehicles intended to be an automated fleet. [9] They committed to Volvo for the purchase of XC90 SUV’s throughout 2019 and 2021. [9] Moving forward there is no information as to what will be Uber’s next move regarding their self driving goals. [9]

    Current Issues with Autonomous Vehicles

    The Dramatic Effect on Insurance

    One area that is being short-sighted among issues surrounding autonomous vehicles is insurance, and the claims surrounding them. The current system in place faults the human driver, however, questions arise when there is no longer a driver present. Insurance companies are now becoming aware of the potential gray areas regarding upcoming vehicle crash legislation where finding who is at fault is easier said than done.

    KPMG has proposed a system of determination for liability regarding autonomous vehicle crashes. [10] Using the level of car autonomy as a guide, they have laid out three separate guidelines for liability. They are assuming the slowing integration of fully autonomous vehicles (level 5), meaning lower levels of autonomy will still be on the roads simultaneously. [10]

    If a car is of 0-2 level of autonomy, meaning a conventional present day vehicle, the fault will remain the driver’s and liability will fall on them. [10]

    If a car is of 3-4 level of autonomy, meaning a car with high automation and minimal driver interference, it is when the driving features were engaged or disengaged at the time of collision. [10] So if there were any auto-driving features engaged, where the driver was not present, then the fault would be the manufacturing company’s, as the driver had not control of destiny. However, if the driver was present at the time of crash (their hands were on the wheel for example), then the fault would be of the drivers.

    If a car is of level 5 autonomy, then the fault and liability would be of the manufacturing company for providing an unsafe vehicle. [10] It being out of the control of the driver for unknowingly riding in a faulty vehicle.

    CMCV.
    KPMG's Proposed Liability Model [10]

    On a grander scale, KPMG expects an 80% reduction in the frequency of auto insurance claims costs, offset by a 40% increase in average claims cost. [10] The increase is due to the higher cost of repairing a complex autonomous vehicle over a present-day vehicle. [10] This likely because of the liability being shifted more frequently to the manufacturer than the driver after collisions. Overall, this results in a 40% decrease in total value of claims across the industry. [10]

    Commercial auto claim value, such as trailers and delivery trucks, is expected to grow by 30%. [10] This is a result of more of them being on the road as logistical companies now have incentive to release more fleets. [10] The savings on previous insurance costs and wages for drivers would then be devoted to running autonomous trucks. However, with more trucks on the road could mean a higher possibility of crashes. Coupled with an assumed higher repair cost than a non-autonomous truck, the increase in total claims value is justified.

    Lastly, product liability is expected to increase by 1200% due to the sheer number of autonomous vehicles on the road. [10] With the early introduction of autonomous vehicles, problems are sure to arise, and much of the liability in these crashes will be on the manufacturers. [10] As a result, the total claims value of product liability will surely rise as the numbers of autonomous cars on the road reaches the millions. [10]

    Insurance Revenue Recovery Opportunities

    CMCV.
    Insurance Loss Across the Industry [11]

    Overall, insurance companies are set to lose out when the majority of drivers switch to autonomous cars. Total revenue from auto insurance premiums is predicted to drop by $25 billion by 2035 due to fewer accidents, decreased insurance pricing, and significantly lower claim frequency. [11]

    To offset these losses, insurance companies can take advantage of three potential revenue opportunities.[11] These opportunities will be made possible during the influx of autonomous vehicles. They are cyber security, product liability, and infrastructure.

    Insuring cybersecurity will essentially protect the user from any criminal hacking and the resulting damages. [11] These include events such vehicle hijackings, unauthorized entry, and ransomware (holding passenger hostage until funds are delivered to hacker). Personal and financial information theft can be insured against as well. [11]

    Product liability insurance will protect the user from software failure across the vehicle. [11] These include software defects, bugs, and connectivity issues with hardware. On the hardware side of the equation, this policy ensures protection against sensory failure and other detection devices. This policy would be crucial in the event of a collision that was caused by aforementioned problems.

    CMCV.
    Potential Recovery Revenue for Insurance Industry [11]

    Due to the future rollout of autonomous vehicles, the infrastructures that arrive with them will need to be insured against, mainly the cloud server systems. The cloud server systems will control traffic flow and vehicle movements in populated areas.[11] As with computer cloud systems, these systems have the potential to be overloaded and can malfunction. Furthermore, the security defences for these systems could potentially be breached. Cautionary autonomous vehicle users may need to protect themselves against such events.

    Insurance will be a topic of high debate and the resulting legislation should to be put in place soon before the impending release of autonomous vehicles. Failing to do so, or having vague regulations could prove disastrous for both government and insurance companies. Additionally, insurance companies will need to take advantage of profitable opportunities following the public release of autonomous vehicles.

    Testing

    Testing has proven to be another issue for autonomous vehicles. Testing on a closed facility is one thing, but testing on public roads has proven to be extremely difficult. There have been several crashes, but none more publicized than the Uber crash in Arizona. [12]

    CMCV.
    Photo of Herzberg and Uber vehicle involved [13]

    The most publicized test result for the autonomous vehicle industry, the crash featured the first death caused by a fully autonomous vehicle. [12] On March 18, 2018, in Tempe, Arizona, a Volvo tested by Uber hit and killed 49-year old Elaine Herzberg. [12] The SUV failed to brake when Herzberg was pushing her bike along a crossway and then came into the path of the incoming vehicle. [12] Investigators found that the car actually saw the woman 6 seconds before impact, but failed to stop. [12] So they turned it off and as a result, Herzberg, who in fact was a "true positive", was never sensed and was hit. [12] They discovered that Uber decided to turn-off the mechanism that emergency stops at close obstacles in front of the vehicle. Uber said that the car was stopping too much, in front of non-dangerous obstacles, calling them “false positives”. [12] So they turned it off and as a result, Herzberg, who in fact was a "true positive", was never sensed and was hit. [12]

    Following the media frenzy, Uber shut down all self-driving testing operations in Arizona. [14] Having been heavily criticized for their role in Herzberg's death, and states such as California showing renewed hesitancy towards future autonomous vehicle testing, Uber finds themselves in a public relation conundrum. [15] This case also provides a precedent for any autonomous car company pursuing public testing. Companies doing future tests on public roads will now have to consider public safety as their top priority as Uber's crash may have now swayed the public opinion of autonomous cars towards negativity and reluctancy.

    Other Autonomous Vehicles

    Autonomous Planes

    Vahana

    Airbus Vahana Flying Taxi [16]

    Autonomy has not been reduced to just the automotive industry. The company A3 of Silicon Valley has already developed a working prototype of an autonomous aircraft. [1] It is single-passenger with dimensions of 6.2m(Width) x 5.7m (Length) x 2.8m (Height) and a weight of 1642 lbs. with a maximum speed of 175 km/h. [1] Its current flying range is 100 kilometres before the lithium batteries start to run out of power, however, these batteries can be replaced in no less than one hour. [1] The total cost per mile is an efficient $2 USD. [1]

    The vehicle takes off vertically and uses its unique “sense-and-avoid” technology allows for improved safety among air traffic and natural hazards including birds and other aircrafts.[2] The system also ensures that the aircraft remains true to its coordinates even after diverting aerial obstacles. [1]

    A3 hopes that the Vahana can become an alternative to short-distance ground transportation such as cars or trucks. [2] The strikingly low operating cost could prove comparable to those transport types and potentially spark interest for Vahana use in other industries. According to A3, Military cargo deployment, EMT services and mountainous search-and-rescue teams are all potential jobs for the Vahana. [2] With a two-passenger prototype set for release in 2020, A3 seems to be on course to realize their goal of a full public release in the near future. [1]

    Autonomous Boats

    Buffalo Automation

    Autonomy for nautical transportation has also started to develop. Buffalo Automation is one of the frontrunner startups involved in the development of automated systems for recreational boats and cargo ships. [3]

    CMCV.
    Buffalo Automation's Automate System's Sensing Features [4]

    Buffalo Automation’s prime area of focus is perfecting their Automate system, which integrates real-time sensor data about waterways with static information to enable self-driving. [3] The system will first receive the coordinates for its “A-to-B” trip. As the ship makes its way to its destination, the system will use its AIS software (automatic identification system) to track and reactively steer away from any ships, debris or people in the water. [3] The sensors and cameras at the head of system can detect up to 24 nautical miles and will still work in turbulent waters. [3] Not just limited to future vessels, the Automate system can be added to existing boats as well, furthering the incentive for shipping companies to add Automate their fleets.

    Buffalo Automation states that the Automate system will improve commercial shipping safety due to early obstacle detection and will save fuel and money as the Automate redirects the ship using only the most efficient routes. [3]

    Perhaps the largest reason for shipping companies to adopt a system like Automate would be the significant cost savings. “Six to seven percent of operating costs is crew,” says Thiru Vikram, the CEO of Buffalo Automation. [3] “Eighty percent of accidents are caused by human error, or fatigue, so we can bring down insurance costs, save fuel, and reduce operating costs.” [3] In 2016, ships transported $1.5 trillion worth of cargo through US ports and the world’s 325 or so deep-sea shipping companies had combined revenue of $10 billion. [3] Taking those lucrative numbers into account, the reduction of insurance, operating costs and fuel would be in the millions. End-of-year bottom lines would likely improve across the industry, making the Automate system an enticing addition for shipping companies worldwide.

    Effects of Autonomous Cars

    On Cities

    With the increasing implementation of autonomous vehicles in our society, our cities will change as advances in the technology make futuristic ideas, such as a flying car, a reality. Three possible ways our cities might change are: the invention of new vehicles, a 3D traffic network and better flow.[5]


    What a driverless world could look like[5]

    Inventing New Vehicles

    The Transit Elevated Bus (TEB)
    China's Transit Elevated Bus[1]

    Space is one thing we need more of, so why are we driving in vehicles with only one person in them 85% of the time?[1] There is currently a debate whether we, as a society, should invest in expanding our mass transit system or allow individual transportation because it's more convenient for travelers.[1] One example of a new vehicle that was invented is China’s Transit Elevated Bus. This bus was elevated to avoid China’s famous traffic congestion and successfully completed its initial run of 300 meters.[2] Unfortunately, the bus project failed because of the unethical business practices of management and illegal funding of the project.[3] Nonetheless, the Transit Elevated Bus is a great example of a new vehicle invention to solve the congestion problems we face today.

    Next
    Next[4]

    Another interesting invention is “Next”. An electronic pod bus concept that was designed by Tommaso Gecchelin, that can connect and disconnect individual pods as needed to transport passengers.[1] This bus solves the problem of transit systems not being able to drop you off at your exact destination, which is why many people prefer to drive individual cars. The inventor also believes that there could even be service pods that can connect to your bus so you don’t have to stop on your way to work.[1]

    3D Traffic Networks

    Currently, our infrastructure is relatively flat.[2] We expand by building new roads and widening current ones, but eventually we will run out of space.[2]

    One way our city might change will be the movement of transportation away from the ground to free up space. Our current problem of running out of room can be solved by building a 3D traffic network.[2] This means inventing new systems that build up rather than out. An example we currently implement of this is a sky train, but there are many more advanced technologies in the works.

    As mentioned in detail before, China’s Transit Elevated Bus is a great example of building upwards. The bus was designed to straddle traffic jams and reduce the number of cars on the road by providing an alternative to get around efficiently.[2]

    CMCV.
    SkyTran X NASA Suspended Magnetic Pods[2]

    Another example of a 3D transportation concept is SkyTran and NASA’s suspended magnetic pods. They are working together to create a transit system, similar to a sky train, but with the added convenience of singular pods so that passengers only have to stop at their destination. SkyTran and NASA hope to launch the project in Tel Aviv, Israel.[2]

    CMCV.
    Pop.Up[3]

    Moving even higher, Airbus is working on a concept they call Pop.Up.[4] This is a flying autonomous vehicle that they hope will replace taxis.[2] The vehicle has three main parts: the land unit, which has the wheels of the autonomous vehicle; the autonomous pod itself which can attach to the land and air units; and the air unit, which allows the pod to fly like a helicopter.[4]

    Better Flow

    Currently, driverless cars are trying to learn our traffic laws and deal with human unpredictability.[2] But what will happen when everything is connected and there is only driverless cars on the road? Cities will have no need for things like red lights, lanes and low speed limits.[2] Cities will be “closer” because it will take less time to get there and you could even accomplish tasks on your way, making you more productive.[2] One implication of this, is that pedestrian and driverless cars may have to be separated, since it would be inefficient if the cars still had to stop for crossings.

    On the Economy

    Universal Basic Income[5]

    Driverless cars and other automated technology will be replacing many of the jobs available in the world we live in today. But what effect does this have on our economy as a whole? Currently, humans work for companies in order to receive an income they use to pay for products from these companies and pay taxes to the government. But if humans are replaced by technology, and therefore no longer have an income to pay either businesses for products or the government taxes, how can the businesses and government sustain themselves?

    Universal Basic Income

    One idea is a universal basic income. The increase of automation “is going to further concentrate wealth in the hands of the few who own the robots".[1] Therefore, “we need to think of a new social structure and new purposes for human beings”.[1] “Jobs in trucking, fast food and factories have long been aveneues for those without higher education to have a middle class living, or simply to stay afloat", but these jobs will be the first to be replaced since they are easy to automate.[1] Universal income works by collecting money from businesses who have displaced humans through automation, and redistributing the wealth to those who have been dispersed, by giving citizens "free money".[1] Although this seems like a solution to the problem, criticism over "taxing innovation" and making "no incentive to work" have been arguments against this solution.[1]

    On Jobs

    Impacts on Other Industries

    Driverless cars will have a massive impact on our society and the way we live our lives. This impact will extend beyond just driving and reach into other industries such as insurance, delivery, and towing.

    ICBC

    One industry that is likely to be affected by driverless vehicles is insurance. Specifically looking at ICBC, there will be no more need for driver’s education courses or tests, thus eliminating the need for the instructors. The possibility of a subscription model for autonomous vehicles might also mean that there would no longer be individual insurance, but instead, company business insurance.

    Delivery

    The second industry that has already started to be effected is delivery. Whether its food delivery, item delivery, or people delivery in the case of taxis, autonomous vehicles are likely to take over. As discussed previously, Ford and Dominos have already partnered up and are testing an autonomous pizza delivery vehicle[2], and Uber’s Otto has teamed with Budweiser to deliver cans of beer over a long-haul trucking route.[3] Regarding the taxi industry, there will be no more drinking and driving, so no need to catch a cab after a night out. Furthermore, communities located outside transit systems will have another option besides taxis, and driving their own car to get to their destination.

    Towing

    The tow truck industry will also be affected by autonomous vehicles. As seen perviously, autonomous vehicles may reduce accidents in up to 90% of cases.[4] This means that the tow truck industry would be called to less accidents, thus decreasing the need for their services.

    Trucking

    Another industry affected by driverless cars could be the trucking industry, as shown by the movie Logan, which takes place in 2029.[5] Interstate highways could be flooded with shipping trucks without drivers. [5]

    CMCV.
    Tesla Semi [6]

    Autonomous trucks bring about their advantages, they can coordinate their movements and create a small platoon to follow each other through long stretches of highways, almost like a train. [7] Thus, improving aerodynamics and reducing fuel costs for their owners. Additionally, truck driving is a high stress job that requires lots of care and attention for hours. An autonomous truck would still require a driver, however, stress would reduce. Furthermore, recruitment would become easier, and shortage of drivers could be eliminated. [7]

    Although, autonomous trucks could also bring some disadvantages. The current technology limits autonomous trucks from being successful in rural areas or in very busy urban areas, a human driver would still be needed to take over if necessary. Safety regulations and cyber security are also major issues that must be tackled before the mainstream adoption of autonomous vehicles, specially if millions of dollars worth of merchandise are being transported. Lastly, loading and unloading duties must still be done by humans. [7]

    Tesla is one of the companies that aims at making this industry much better by introducing its Tesla Semi. This can potentially be one of the first steps before a 2029 full of automatic trucks, as seen in Logan. [5] The Tesla Semi includes multiple safety and efficiency features, as well as a low cost of ownership.[8] Tesla already has over 2000 preorders for the Semi, some notable companies that pre ordered are Pepsico with 100, Sysco with 50 and Budweiser with 40 trucks. [9][10][11]

    Looking into the Future

    Cars as Businesses

    If in the future you can own a driverless vehicle, why not send it off to make money while you’re at work? Your car could pick you up from your house, drop you off at work and then leave again to pick up and drop off other passengers until you need to be picked up. Although this seems like a good idea, it may face challenges in the future from automakers such as Tesla. Tesla has already banned the use of its vehicles with services such as Lyft and Uber.[12] But all hope is not lost, as Tesla announced that they are starting to work on their own ride sharing service for their cars.[12]

    Subscriptions

    Most car owners only use their vehicle 5-10% of the time in a day.[12] So why bother owning a car? In the future, there may be subscription apps that will let you hail autonomous vehicles similar to Uber or Lyft's services. You could use the app to hail a car or pre-program one to pick you up every morning at the same time to go to work, but would not have to incur the costs and space associated with maintenance and storage of the vehicle.

    Delivery

    As autonomous vehicles take over regular cars, they will likely replace delivery drivers. Whether it is food or item delivery, there are companies already starting to prepare for the change.


    Ford X Domino’s Pizza Delivery[2]

    Ford has partnered with Domino’s pizza to test their driverless vehicle technology while delivering pizzas in their test city.[1] The driverless vehicle can even find you if you are not at an address, but instead, are simply walking down the street.[1] Once the car arrives, a panel on the window asks you the last four digits of your phone number and if correct, will open the window to allow you to grab your pizza from the car.[1] Pizza Hut has also teamed up with Toyota to create their fully automated delivery vehicle, which could also be used as a vehicle to cook pizzas in the future. They estimate to begin testing in 2020.[2]


    Uber’s Otto X Budweiser Autonomous Truck[3]

    Since long-haul commercial trucks are responsible for delivering 70% of goods in America, there is a strong incentive to automate this process to make it more efficient.[1] There is a great potential to save time because unlike humans, autonomous vehicles wouldn’t need to take breaks to rest on a long trucking trip. Uber’s Otto partnered with Budweiser in 2016 and successfully completed a 120 mile drive safely to deliver Budweiser.[1]

    Organ Donors

    The driverless car movement is changing the way we live, but it may also change the way we die. Roughly 6,500 Americans die waiting for an organ transplant each year, and another 4,000 are removed from the waiting list because they are deemed too sick for a transplant. [2] One of the most reliable sources for healthy organs and tissues are the more than 35,000 people killed each year on American roads. [2] Currently, 1 in 5 organ donations comes from the victim of a vehicular accident, which is why departments of motor vehicles ask drivers whether they want to be donors. [2] 94% of motor vehicle accidents are due to driver error and with that high of a percentage, it is not hard to see how driverless cars could affect organ donations.[2] Of course, the driverless car movement’s goal is to reduce the number of driver error accidents and save more lives. However, as autonomous technology moves at an extremely rapid pace, it is important to consider that other necessary technologies, such as artificial organs, might not be able to keep up.

    Summary

    An autonomous vehicle is a vehicle that is capable of navigating without human conduction by sensing its environment. [3] These vehicles have become a prominent reality and have begun to pave the way for future systems in which human input is no longer needed and computers are able to take over the art of driving. Specifically, driverless cars are a type of autonomous vehicle that have rapidly began to dominate the autonomous movement.[3]

    There are different levels of autonomy for driverless vehicles, and big companies such as Tesla, Google, and Uber are fighting to be the top competitor with the safest AI on the market. These companies spend billions of dollars researching, developing, and testing their autonomous systems and also invest heavily to make improvements in their cars autonomous parts. Autonomous vehicles sense their environment by using specialized components such as image sensors, radar sensors, lidar sensors and the cloud. As the vehicles become safer, the more they will be used in the real world, and this brings up concerns such as safety, regulation, pricing, liability and concerns of hacking potential.

    Autonomous vehicles will not only change the way we drive, but also how the world looks around us. Our cities will change, there will be new autonomous vehicles invented, and industries will disappear. Looking into the future, AI technology such as automated vehicles, may replace many of the jobs, leading to mass unemployment. One solution that has been debated when the “robots take over”, is a universal income where companies who have replaced humans, will have to pay into a pool that will support the displaced citizens.[4] Another interesting future implication of autonomous vehicles specifically, is organ donations. Currently, 1 in 5 organ donations comes from the victim of a vehicle accident. [2]

    All in all, it looks like autonomous vehicles are here to stay, but debates over the concerns of liability and ethical implications of autonomous vehicle coding are ongoing. It is likely the autonomous vehicle technology will be ready before our laws are.

    Authors

    Allison Kincaid Andres Chavarriaga Annika Ng Travis Uyesugi
    akincaid@sfu.ca andresc@sfu.ca ayn3@sfu.ca tuyesugi@sfu.ca
    Beedie School of Business
    Simon Fraser University
    Burnaby, BC, Canada
    Beedie School of Business
    Simon Fraser University
    Burnaby, BC, Canada
    Beedie School of Business
    Simon Fraser University
    Burnaby, BC, Canada
    Beedie School of Business
    Simon Fraser University
    Burnaby, BC, Canada

    References

    1. 1.0 1.1 https://www.youtube.com/watch?v=sIlCR4eG8_o Retrieved on June 14, 2018
    2. 2.0 2.1 2.2 2.3 2.4 http://www.slate.com/articles/technology/future_tense/2016/12/self_driving_cars_will_exacerbate_organ_shortages.html Retrieved on July 2, 2018
    3. 3.0 3.1 https://www.techopedia.com/definition/30056/autonomous-car Retrieved on June 2, 2018
    4. https://www.youtube.com/watch?v=o0pqSKfn8FU/Retrieved on August 01, 2018
  • Personal tools