Autonomous Transportation - D100

From New Media Business Blog

Jump to: navigation, search

Autonomous Transportation



There’s no arguing that autonomous technology is one of the most exciting and promising developments in transportation. Automobile and tech companies are racing one another to become the first to reach a new level of autonomy or release a new groundbreaking feature. More and more companies are seeking permits to test their vehicles in the public and autonomous transit services are already available in some places.

Autonomous vehicles come in different levels, but at its core, these vehicles are designed to transport someone from point A to point B. Companies are striving to achieve this through the use of specialized software and hardware. Most autonomous vehicles (especially the forerunners in the market) implement a combination of radar, LIDAR, and video cameras. Such sensors provide the software with awareness of the vehicle’s surroundings and the software will determine how to navigate the vehicle [1].

In an ideal world, fully autonomous vehicles will not only create safer roads, but it will also increase the efficiency of our day to day travels. Although the technology may be getting close, the amount of trials that must be conducted, the strict and different region to region regulations, and the fatal accidents involving autonomous vehicles all act as speed bumps that slow the arrival of fully autonomous vehicles [2].

The Journey To Autonomous Transportation

Vancouver SkyTrain

Vancouver SkyTrain (1985)

After attempting to deal with traffic and transportation issues in the city for several years since the end of streetcar service, the government chose to adopt Advanced Light Rail Transit, a new Canadian technology. Completed in time for the 1986 Expo, the SkyTrain was the first driverless urban rail system and served as the centrepiece of the Expo [3].

Park Shuttle (1997)

Operational since 1999, ParkShuttle is the first driverless vehicle system in the world [4]. It used artificial reference points embedded in the road to verify its position. These autonomous vehicles did not have steering wheels, pedals, nor did they have a safety driver on board.

The first pilot project to launch was in 1997, at the Schiphol Airport in the Netherlands [4]. The system transports over 2000 passengers each day and the people that have used ParkShuttle noted that the system was much more reliable than traditional buses.

With a completely automated system, consistent wait times and schedules become much more manageable.

VisLab Intercontinental Autonomous Challenge (2010)

In 2010, Italy’s VisLab ran the VisLab Intercontinental Autonomous Challenge (VIAC), marking the longest trip completed by an autonomous vehicle. Four autonomous vehicles set out on a three-month long trip, starting from Parma, Italy on July 20th, 2010 and ending at the Shanghai Expo in China on October 28th, 2010 [5].

Navia (2014)

In 2014, Induct Technology started selling Navia, their golf cart/minibus-looking autonomous vehicle. Navia was the world's first commercially available self-driving vehicle. It can carry up to 8 people and can travel just over 20km/hr. This 100% electric vehicle travels using laser-based sensors and charges by induction (using magnetic fields) [6].

Much like the ParkShuttle, Navia can be instructed to travel a specific route or Navia can be summoned via app, in which the user can input their destination onboard.

Delphi Automotive (2015)

Only 5 years after Vislab’s autonomous vehicle managed to travel 15,900 kms in 3 months, Delphi Automotive produced a vehicle which was able to do 5,500 km’s in only 9 days[7].

From the Golden Gate Bridge to Manhattan, an Audi SQ5 equipped with Delphi Automotive’s autonomous driving technologies completed the first coast-to-coast trip ever taken by an autonomous vehicle [7].

nuTonomy (2016)

NuTonomy launched the world’s first driverless taxi service in one of Singapore’s business districts, inviting a limited number of individuals to download their app and call their autonomous taxis free of charge [8].

While there was an engineer behind the wheels of the modified vehicles at all times, these road trials allowed the team behind the cars to gather vital data which will help them further improve their software and vehicles[8].

Audi (2017)

In 2018, the 2019 Audi A8 became the first production car to reach level 3 autonomy. They came with a feature called “traffic jam pilot”, which was capable of taking control of the car in slow moving traffic (traffic jams) on divided highways. Users stuck in these constantly stopping and starting jams can enable the feature and the vehicle will completely control the steering, braking, and accelerating, allowing the driver to be engaged in another activity completely [9].

While the actual availability of the feature will depend on the laws and regulations of different markets, it was still the first production level feature that had reached this level of autonomy.

Aptiv & Lyft's taxi service in Vegas.

Aptiv (2018-2019)

Last year, in January, Aptiv and Lyft launched a pilot project to test a robo taxi service in Vegas which was only supposed to last for one week. The program is still going strong to this date and as of last week, Aptiv revealed that the company has given over 50,000 rides [10].

Aptiv’s program also shed some positive light on the public opinion on self-driving vehicles. While it is by no means a proper representation of the general public, out of those used Aptiv’s robo taxi service, 92% felt “extremely safe” and the rides averaged 4.97 stars out of 5.

Present Day Leaders

Many companies have promised delivery of autonomous vehicles and most have invested significant resources in its development; however, most companies have not been able to meet their deadlines. In recent years, these large companies have admitted that the production of a fully autonomous vehicles would be more difficult than they had originally anticipated.


As of now, Tesla is the only company that has not postponed their launch date, as Elon Musk confidently announces to the world that he will launch the first fully autonomous vehicles by the end of 2021.

Earlier in 2019, Elon Musk announced that all Tesla vehicles being produced would have the hardware required for full self-driving and that the only missing piece would be the software [11].Tesla has also began working on their next-generation chip, which Musk believes will be able to provide more sophisticated self-driving capabilities.


In late July 2019, Waymo announced their collaboration with DeepMind. With DeepMind’s neural network and AI technology, Waymo is hoping to find ways to improve their self-driving algorithms and learning processes [12].

Just a month earlier, Waymo (partnered with Lyft) opened several self-driving minivans to the public in a small part of Phoenix, Arizona [13].When users hail a ride with Lyft in the specific area, they will have a choice of selecting a regular vehicle, or the self-driving minivans, which all have a human driver behind the wheel. These real life trials will ideally give Waymo very valuable data that will help them further develop their software.


After one of Uber’s self driven vehicles struck and killed a pedestrian in 2018, Uber’s autonomous vehicle program faced many questions and CEO Dara Khosrowshahi even reportedly considered terminating the program. The latter never happened and in June 2019, Uber announced the third generation of their self driving vehicles. Learning from their mistakes, the new version has been designed to have many levels of redundancy when it comes to back up systems. On failure of any primary systems, the vehicles will come to a complete stop [14].

The Technology

Levels of Automation

Autonomous vehicles are held to a hierarchy-based schema that defines each level of automation. A multi-part continuum suggested by the National Highway Traffic Safety Administration (NHTSA) and Society of Automotive Engineers (SAE) [15] [16] describes each level:

Level 0 (no automation):
This is the stage at which most vehicles on the road are currently at. The human driver is responsible for the performance of all tasks, even if some may be enhanced by active safety systems, such as automatic braking.

Levels 1 and 2 are considered to be advanced driver assistance systems (ADAS).

Level 1 (driver assistance automation):
Vehicles at this level have one or more specific automated functions that operate independently of each other. For example, while the driver still has overall control, the vehicle can perform sustained control of either steering or acceleration and deceleration.

Level 2 (partial automation):
There is the automation of at least two primary control functions meant to work in unison. Although it is no longer necessary for the driver to be continuously hands-on, full attention to monitoring the driving environment is required at all times.

Levels 3, 4 and 5 are considered to be automated driving systems (ADS).

Level 3 (conditional automation):
At this level, the vehicle’s automated driving system is able to perform all driving tasks under certain circumstances, and human intervention may still be expected on short notice. In all other circumstances, a human driver is expected to handle the task of driving.

Level 4 (high automation):
Automation at this level allows human drivers to pass full control to the vehicle, but only in limited conditions and environments, (eg. dedicated lanes or zones). Human intervention is optional.

Level 5 (full automation):
The vehicle can perform all safety-critical driving functions and monitor roadway conditions in all environments and settings. Human intervention is not anticipated at any time throughout the trip, and safe operation relies solely on the automated vehicle system.

Levels of automation.

Sensor Systems

LiDAR, or light detection, is considered the most reliable and robust sensor technology. LiDAR uses laser range finders, which emit light beams, to calculate the time-of-flight until a reflection is returned by an object in the environment as a measure of distance to obstacles [15]. A key limitation is its range. For obstacles further away, detection is less optimal. LiDAR is also less effective when detecting objects made of poorly reflective materials, such as pedestrians.

For example, in mines, LiDAR can be used to generate clearer images of the underground terrain, and detect instances of cave instability or compromised infrastructure [17].

Radars also help with detection and ranging using high-frequency radio waves to detect obstacle velocity, directions, and distances. However, it is even worse than LiDAR when dealing with objects of poor reflectivity.

Cameras identify objects using visible light, but have obvious environmental and conditional disadvantages depending on the darkness, brightness, and weather conditions of surroundings.

Sensor suites, such as Global Positioning Systems (GPS), are also useful for localization and determining the vehicle’s current geographical position through the communication of satellite signals [18].


For autonomous vehicles to work efficiently, it is crucial that cars get real-time data of the latest traffic, weather, surface conditions, construction, adjacent cars, road infrastructure [19]. This information gives the vehicle understanding of its surrounding operating environment, allowing the system to better anticipate braking or to avoid hazardous conditions.

Connectivity and communication rely heavily on technologies such as Dedicated Short-Range Communications, 5G, Vehicle to Vehicle communication, and Vehicle to Infrastructure communication.

Dedicated short-range communications (DSRC) is two-way communication in a 5.9GHz band allowing high transmission rates of data over a moderate range [20]. DSRC gives the vehicle the capability to send drivers real-time alerts and opens up opportunities for V2I and V2V communications.

5G is another communication technology that is nearing completion. 5G will allow for higher capacity and better coverage of longer distances with less latency, in comparison to DSRC, supporting device-to-device communications with improved reliability [21].

[22] Vehicle to Vehicle Communication

Vehicle to Vehicle Communication V2V or Vehicle to Vehicle Communication is a crash avoidance system that relies on sensors, wireless networks and dedicated short-range communications (DSRC) to allow vehicles to send messages and information to each other. DSRC can be comparable to wifi, with a range of up to 300 meters, 1000 feet, or approximately 10 seconds on the highway [22]. The data sent from one vehicle to another usually includes speed, location, direction of travel, braking, and loss of stability.

For example, if the tail emergency braking lights of a car in the front is obscured to the driver in the back, V2V is especially useful. The technology manages blind spot detection at left turns and intersections, areas with the highest crash :incident rates, and help to raise driver awareness during lane changes.

Overall, V2V implementation is projected to prevent of 9,300 deaths annually, and reduce total automotive-related deaths by 25% [21].

One of the earlier developed components of V2V is the Automatic Emergency Braking (AEB) system, which has been available on some US car models since 2006.

While much of the technology is still in development, some models on the market have already adopted V2V safety features. Cadillac’s 2017 CTS Sedans are one of the first cars to be V2V-equipped, sharing information from up to 1,000 feet :away to scan the vicinity for other vehicles, track their positions, directions and speeds, and warn drivers of potential hazards that may otherwise be unseen [23].

Vehicle to Infrastructure Communication V2I or Vehicle to Infrastructure communication is the wireless (and typically bi-directional) exchange of data between vehicles and road or nearby infrastructure enabled by a system of hardware, software, firmware [24]. Infrastructure components include advanced road and lane markings, road signs, and traffic lights to provide info to the vehicle and vice versa.

A prime example is having road markings that would be clearly visible to vehicle sensors under any weather conditions, such as rain or snow.

Or alternatively, suppose there is a car passing through a construction site, and a worker waves the car by. How does the autonomous vehicle know that the worker is an authoritative figure? We could have a chip planted on the worker that :signals to the car that they have the right to override, for example, a red light [25].

Lastly, if autonomous vehicles begin using high-occupancy-vehicle (HOV) lanes, V2I would be useful to let the car know to switch into autonomous mode when entering the lane [25]. Simultaneously the infrastructure should also check to ensure that the vehicle is in the correct mode, emphasizing the importance of two-way communication.

Software and Control Algorithms

Neural networks and rule-based decision making [26], help to reliably capture data (perception stage) from sensors and communications technologies. It’s essential that the software is robust, that is, capable of handling exceptions to input data, and fault-tolerant in order for the system to respond appropriately.

Another main role of this component is to interpret sensor data, such as lane markings from images of the road, behaviour of other vehicles from radar data, and use that data to make plans about the vehicle’s own actions (planning stage). Actions include the trajectory down the road and immediate decisions such as accelerating and/or changing directions (control stage).

The complex algorithms evaluate, compare, select and execute best action from a number of maneuvers in time-sensitive situations, taking into account the vehicle's speed, obstacle trajectory, position/behaviour of other vehicles etc.

Benefits and Risks of Autonomous Transportation


As we move towards a future in which fully autonomous vehicles are becoming closer to reality, the revolutionary effects they would have on society and the technological landscape are becoming more and more clear. From improved public safety to efficiency and time saved, autonomous vehicles will bring about many advantages and benefits.

Increased Safety and Reduced Accidents

From distracted drivers to driving under the influence, it is apparent that most road related fatalities can be attributed to human error. The 2016 Summary of Motor Vehicle Crashes states that “there were 10,497 alcohol impaired-driving fatalities, representing an average of one alcohol-impaired driving fatality every 51 minutes; [additionally], an average of 102 people died each day (one every fourteen minutes) in motor vehicle crashes” [27]. "Self-driving cars have the potential in the future to reduce deaths and injuries from car crashes, particularly those that result from driver distraction," said House Energy and Commerce Committee Ranking Member Frank Pallone, Jr. (D-NJ) [28].

Reduced Traffic Congestion

With the near limitless resources available to Alphabet Inc., parent company of Google, it is safe to say that compared to a driver’s instinct and judgement, Google Maps will usually be better informed when it comes to the quickest and most efficient route to your destination. Similarly, driverless cars use maps and sensor information to objectively determine locations and the best routes to avoid traffic; therefore, commuting time will be reduced, as driverless cars will reduce traffic [29].

Additionally, human drivers naturally create a stop-and-go pace in traffic, such as when a driver decides to merge or lane change. Cornell University launched a field experiment that suggested, “flow control will be possible via a few mobile actuators (less than 5%) long before a majority of vehicles have autonomous capabilities” [30]. In other words, when they controlled the pace of autonomous vehicles within their field studies, the autonomous vehicles controlled the traffic flow by dissipating the stop-and-go waves; concluding that autonomous vehicles can revolutionize the control of traffic flow [30].

Reduced CO2 Emissions

A reduction in traffic congestion will be subsequently followed by a reduction in CO2 emissions.

Parking and Parking Structures

Once driverless vehicles take over the roads, parking spots will no longer need to be as large as they are today. Fewer parking areas will be needed since parking spaces can be made 15% smaller since cars will no longer need to leave space for passengers to get out [31].

Accessible Transportation for All

Due to safety regulations, many seniors and persons with disabilities cannot drive. Autonomous vehicles can provide these individuals with access to the open road. Additionally, this technology can offer these individuals more than just transportation, but a life with more independence.

During the full committee markup of the SELF DRIVING Act, Digital Commerce and Consumer Protection Subcommittee Vice Chairman Representative Gregg Harper (R-MS) strongly stated that self-driving technology can break down the transportation barriers facing the disabled community: “With self-driving cars, tasks like commuting to work, going to the doctor, and visiting family across town could become easier for seniors and those with disabilities” [32].

Autonomous vehicles will play an immense role in the future by improving the lives of many individuals who live with mobility and accessibility challenges.

Convenience and Time Saved

“During 2016 and 2017, the American Driving Survey found that on average, drivers spent 51 minutes driving approximately 31.5 miles each day, making an average of 2.2 driving trips” [33]. Based on these statistics, it is noteworthy that autonomous vehicles have the potential to free up approximately two or more hours a day for commuters.

Additionally, many individuals choose to take the bus, train or a taxi to save time. The driverless car will mimic the process of a taxi, bus or train, allowing individual to use their time as they would typically in a vehicle that does not require their full attention [34].


As we get into higher levels with the technology, there come greater security risks, with unknown hacking potential. These windows of risk are cause for consumer apprehension. A 2017 survey by the insurance company, American International Group, revealed that 41% of consumers have reservations about self-driving cars [35].

[36] Object Identification with Machine Learning

Machine Errors

Today’s autonomous systems rely heavily on machine learning. The algorithms and software programs are fed datasets of behaviours and images, which help them learn to classify and respond to different situations. Inputs include thousands of images of potential obstacles, including pedestrians, bicyclists, skateboarders, or ambulances, etc. [37]. Eventually, the machine is expected to recognize these objects on its own.

Since learning is so highly dependent on the inputs, the greatest challenge is finding images representative of every possible situation and edge case for the machine to learn from, and in turn, be able to anticipate the object’s movements. Classifications made by the autonomous system may not be accurate, and even then, not quick enough.

In a 2017 case, where a Uber vehicle collided with and killed a woman walking her bicycle in Tempe, Arizona, it was reported that the vehicle did, in fact, detect and see the pedestrian. The victim was first identified by the vehicle as an unknown object, then as a vehicle, then a bicycle, re-adjusting its expectation of her trajectory following each classification [37]. Only about a second before the impact did the self-driving system decide that emergency braking was required to mitigate a collision. In this case, human intervention was needed, but the driver did not have the time necessary to respond.

Hacking Potential

As autonomous vehicles are so highly dependent on the system and connectivity to perform the different tasks, there is an increased risk for breach of cybersecurity. Currently, approximately 84% of security professionals and and auto-engineers are worried that automakers are not keeping up with the rapidly changing security threats in a recent 2019 survey by data protection research group Ponemon Institute [38]. Consumers have also expressed concerns for autonomous vehicles. The American Automobile Association found in a 2018 consumer study that three in four (73%) of the 1,008 surveyed Americans were afraid of self-driving cars, and only 19% felt comfortable about self-driving cars transporting their children or loved ones [39].

Connectivity through DSRC, wifi, and soon, 5G, opens up opportunities for unauthorized remote access to the vehicle network, which can have costly and hazardous effects to users of autonomous vehicles and the surrounding environment. Not only will personal information be at risk of being stolen for misuse, users’ physical safety will also be jeopardized when control is seized from the safety-critical systems by malicious hackers.

In 2015, two researchers, Charlie Miller and Chris Valasek challenged the security and multimedia systems of a Jeep Cherokee, and successfully took remote control of every component of the vehicle, from steering, engine, transmission, braking system, windscreen wipers, air conditioner and door locks. This was done simply by hacking into the Sprint cellular internal network and performing a mass scan of IP addresses, giving the two access to all Chrysler cars equipped with the multimedia system [40]. As a result, Fiat Chrysler recalled over 1.4 million remotely hackable vehicles [41].

Additionally, according to Skanda Vivek, a postdoctoral researcher at the Georgia Institute of Technology, even if only a small number of internet-connected self-driving cars on roads are hacked, the flow of traffic would be completely frozen, stalling the travel of emergency vehicles such as ambulances [42].

While these risks are growing concerns as autonomous vehicles of higher levels approach commercialization, technologies are in development to prevent hacking and to build stronger security systems. For example, SK Telecom has recently announced the launch of a potential solution based on Quantum Encryption which involves the installation of an integrated security device to protect various electronic units and networks within the vehicle [42]. By using a quantum random number generator and Quantum Key in transferring the vehicle’s data, the technology will be used to fundamentally prevent hacking.

Autonomous Trucking

In Canada alone, trucking is a $65 billion industry. We rely heavily on this industry for efficient and sustainable freight transportation of products. The Canadian Trucking Alliance (CTA) of Canada statistics show that 90% of all consumer products and foodstuffs were shipped by truck throughout Canada, and about two-thirds in the US [43].

A study by the CTA also found that the current trucking industry will be short as many as 48,000 drivers by 2024 [44]. Currently, the average age of the Canadian truck driver is 48.


There are three main areas where autonomous trucking can be implemented.

Repetitive Routes

Industries that require trucks to travel defined routes, at low speeds, and of small sizes, will be some of the first to allow for full autonomy trucking. Examples include the garbage collection and mining industry.

Volvo’s fully autonomous truck is the first in the world to be tested in operations in the Kristineberg Mine, traveling a distance of 7 kilometers of narrow mine tunnels, 1,320 meters underground [45]. This initiative is part of a research product dedicated to improving safety and productivity in areas such as mines, ports, and other restricted and controlled environments that demand for a lot of repetitive driving. Sensors, including GPS, radar, and LiDAR are fitted onto the Volvo truck to continuously monitor its surroundings, allowing the system to create a map of the mine’s geometry and to create a route through the tunnels for the truck to follow [46]. The map also informs the system of steering, gear changes, and speed.

Highway Driving

Another one of the first uses of autonomous trucks is to handle highway driving. While the human driver will still have control when dealing with complex environments and conditions such as city driving, bad rural roads, or loading, the AI system automatically takes over when the vehicle gets on the highway, allowing the driver to disengage and rest.

This is especially beneficial to long-haul drivers who can at times drive up to 13 hours a day or 70 hours a week [47]. Being able to rest during long drives can help reduce the stress of the job. Additionally, vehicles will no longer need to stand idle when drivers are taking mandatory breaks. Under ideal conditions, one driver and an autonomous truck can replace the work of three current drivers.


[48] Peloton Platooning

Platooning is the linking of multiple trucks to drive in unison at a close distance to reduce aerodynamic drag. The leading truck acts as the head, with the vehicles following behind reacting and adapting to the changes in the head’s movement [49]. Fuel usage is reduced when posterior trucks draft off the leading truck.

Peloton, a Silicon Valley company currently developing this technology, has introduced a Level 1 system called PlatoonPro, onto the market. In this platooning system, only two trucks are linked. The driver in the head truck steers, accelerates, and brakes as normal. The truck following also requires a human driver, but only for steering. Its speed is controlled by the first truck to maintain a close distance between the two [50].

In past testing done for Peloton, it was found that the fuel use for the lead truck was reduced by 4.5% and following trucks by 10% [51]. The findings were consistent with other past trials.

Depending on the level of autonomy, the entire chain of trucks can be completely run by the autonomous system, or at lower levels, the leading truck may still be fully controlled by a human driver. Due to human error and the close distances required, platooning of the rear trucks is too dangerous to be entirely handled by human drivers, with risks of crashing or pile-up.

Platooning is an instance where Vehicle to Vehicle (V2V) technology is crucial. Human reaction time behind the wheel if the car in front brakes is approximately 1.5 seconds [52], but using V2V transmission, the signal can be communicated in only 0.1 seconds [53].

This application combines safety and collision mitigation technologies with vehicle communications and automated vehicle controls to tether trucks together in formation. Adaptive Cruise Control and Automated Manual Transmissions are enough to enable truck platooning, although more advanced systems will also require things like Automated Lane-Keeping, a system that automatically adjusts the lateral direction of the truck if it moves out of line of the road lane markings.

Costs of Platooning

The cost of implementation depends on the desired level of autonomy. Specifically, V2V communication hardware and equipment is projected to range from $100 - $350, and a simple system such as Blind Spot detection will range from $250 - $850 [54].

More advanced systems such as Adaptive Cruise Control could range around $1500.

In a 2016 study by the consulting firm Roland Berger, platooning as a whole could range from $1,800 at Level 1 to just over $23,000 at Level 5 per truck [54]. The differences in prices between the levels is due to the increased number of sensors needed and other built-in safety systems, as well as R&D for technological advancements that have not been fully developed yet. Of course, these are just rough estimates, and other studies do cite different costs.

The discrepancies in costs could be explained by evaluations of a single prototype versus under the assumption of some level of market adoption and achieving economies of scale.

Pros & Cons

Potential benefits include additional road capacity and on-road safety, with reduced collisions and fatalities. However, if the drivers are too quick to trust the autonomous system, say at Level 3, they may be unable to take back control in a timely manner when necessary, mitigating some of the safety benefits.

Autonomous trucking will improve the current fuel economy and reduce emissions, but the extent depends on a fleet-to-fleet basis. Early studies have found that approximately 65% of current long-haul truck miles could be platooned, with the potential to reduce total truck fuel consumption by 4% [55].

Despite this advantage, faster travel could undermine the fuel benefits of autonomous trucking technology. In cases where trucks are platooned too closely to each other, there will be significant costs to reduce engine cooling. The additional requirement for cooling fans could possibly nullify some benefits of fuel savings.

For truck drivers, the technology will make driving easier and increase the operational efficiency of businesses. However, implementation can be challenging and may require regulatory changes to hours of service. Platooning can also increase stress and have negative health effects, especially for rear truck drivers having to drive at such a close range to the other trucks.

For fleet owners, driver wages are the second largest expense behind fuel costs, with wages and benefits making up nearly 35% of average marginal costs per mile [56]. While autonomous trucking technology reduces labour costs, it will be replacing millions of jobs. As the technology is introduced into the market, it will undoubtedly challenge the issue of truck driver shortage in Canada, but the extent is difficult to predict.

Although many companies are predicted to be replacing truck drivers with autonomous vehicles, some businesses are taking a slightly different approach. Starsky Robotics, for example, sees value in the human-driven trucking piece of operations. Even as self-driving trucks are adopted into the business, trained drivers have the opportunity to be promoted to remote operators who will be responsible for taking over and navigating the trucks remotely from an office space to ensure smooth travel [57]. Compared to the traditional work, these employees will have a much better work-life balance, being able to spend less time on the road and more time with their families at the end of the day.

With the commercialization of autonomous trucking also comes the creation of new jobs for the design, building, and maintenance of these vehicles [58]. Given the aging workforce, there is a chance that by the time autonomous trucks are brought to large-scale commercialization, the current workforce will have already been phased out. With growing trends in technology development, the new workforce may evolve with the cultural shift, actively seeking more skilled and technical job positions as opposed to truck driving work. However, those looking for careers in driving may find a new place in the delivery industry, with the rapid growth and adoption of e-commerce.

Autonomous Boats

As the technology behind autonomous transportation has been improving exponentially over the past few years, new applications have begun to come to fruition. Autonomous Boats are a prime example of an alternative application that offers high potential benefits and has begun to attract significant investment.

The benefits of this technology in this industry are significant, currently 90% of global shipping is by sea which represents a multi-trillion dollar industry [59]. Even marginal cost savings in this industry can potentially represent billions in savings alone. Currently the technology is in the early phases of development as multiple companies have begun sea trials for the viability of the technology.



It has been estimated that human error can be attributed to approximately 62% of shipping accidents making it the most significant and avoidable cause of shipping accidents [60]. These accidents have been well documented in cases such as the Exxon Valdez accident in Alaska which cost over $7 Billion through the subsequent environmental damage [61].Through the use of Autonomous boat technology, these accidents can be largely avoided due to the fact that the technology removes the need for human operators. These ships will be able to respond to threats much faster than their human counterparts and constantly remain aware of their surroundings. However, improvising to new situations unaccounted for in the code development is an advantage that human operators will continue to have over autonomous boats.


In addition to the cost savings through the reduction of human error accidents, Autonomous boats offers a variety of other cost savings benefits. Significant cost savings would come from reduced fuel consumption. Autonomous ships would be able to take the most efficient path and conduct “just-in-time” shipping [59]. The most efficient path refers to boats taking adjust paths that take into account weather patterns, this path then would be the best fuel efficient alternative instead of simply the most direct path. Just-in-time shipping refers to boats that would be able to depart their destination and arrive exactly when they are scheduled. This method of shipping could potentially reduce costly time spent waiting outside the harbour while other ships are unloaded or loaded.

Another significant source for cost savings would be attributed to the changes to the ship design. Autonomous ships would be able to save 6-7% of operating costs without a crew and furthermore be able to carry more cargo without space needed for crew quarters and amenities [62]. These simplifications in the ship’s design will also significantly change the look of ships in the future, with autonomous ships becoming more streamlined.

[63] Rolls-Royce

Overall these improvements to the ships performance and efficiency make this technology a worthwhile investment for both shipping companies and manufacturers.

Lower Risk of Piracy

A lower risk of piracy is another added benefit of autonomous boat technology. In 2018 alone, piracy accounted for over $1 Billion in economic damage [64], these have been primarily attributed to kidnappings as well as the wider route ships take to avoid more dangerous waters. Autonomous boats who operate without a crew would be able to avoid the risks of kidnappings and thus be able to take the most fuel efficient route through waters with active pirates. Additionally without active controls on board, the risks of boat hijackings can be reduced.

Risks and Challenges


Certification represents one of the greatest challenges to making this technology a reality. The International Maritime Organization (IMO) tasked with the “safety and security of international shipping” [65] have only begun in their evaluations of the technology. Similar to the roadblocks for autonomous cars, numerous variables encountered in the water must be considered and addressed by the technology before it can be certified as safe.

Examples of this include:

- Interaction with ports
- Interaction with other boats
- Response to weather conditions
- Response to marine wildlife
- and many more

Currently autonomous boats have not yet been certified for use in international waters.

Technology Hacks and Malfunctions

Due to the heavy reliance on technology, autonomous boats are susceptible to issues such as being hacked and technology malfunctions. The consequences of these are very high and this presents a significant challenge to those developing the technology today, this subsequently increases the difficulty of the certification process for autonomous boats. As the technology is only in its infancy, creating secure systems for autonomous boats are still in the early developmental stages.


Global Positioning Systems (GPS), LiDAR, and Thermal Cameras are the technologies powering autonomous boats. These technologies would be employed across hundreds of cameras across the boat to detect incoming boats, marine life, ports of entry and help the vessel navigate through weather [62].The GPS operates to help the ship navigate and provides its constant location to control teams back onshore. LiDAR operates very similarly to the functions on an autonomous vehicle where the technology is used to sense other objects and the distance between them and the ship. Thermal cameras help spot marine life in the water that would be normally very difficult to detect by traditional sensors.

Present Day Leaders: Autonomous Boating

Rolls Royce is one of the leaders in developing autonomous boat technology today through their Advanced Autonomous Waterborne Applications Initiative in partnership with the Finnish Agency for Technology [66]. Currently the organization has begun development of remote guided semi-autonomous boats. These boats would be remotely controlled through a global control center removing the need for a crew onboard. In addition to this, the company has begun to develop concepts for IOT technology onboard for continual equipment monitoring. Remote controlled boats are only the 1st step in the company’s timeline towards truly autonomous boats with a timeline for fully autonomous boats ready for launch in 2035.

[67] Autonomous ships: The Next Step

[68] KongsBerg + Yara

This project is a collaboration between the companies KongsBerg and Yara with a planned 2020 target for a fully autonomous boat to be completed. In addition to the autonomous technologies, this project is aiming for their vessels to be fully electrically powered [66]. Currently testing has begun with short deliveries being completed in Norway fjords to test the feasibility of the technology. This project however has faced numerous delays and it is unknown whether this project will be completed on time.

Alternative Uses for Autonomous Boats

[69] Mobile Infrastructure

Mobile Infrastructure

The Massachusetts Institute of Technology and Amsterdam’s AMS Institute have partnered up to create on demand infrastructure involving specialized autonomous boats that move together to form temporary bridges [69]. The advantages of this are mobile infrastructure platforms that can be formed on demand to alleviate city congestion. Applications such as this open up possibilities for autonomous boats to impact far more than the shipping industry.

Ocean Clean-up

Ocean clean-up is another impactful application for autonomous boats. Ranmarine is a great example of a company applying this technology, pictured below is their product waste shark which autonomously moves through the water collecting trash [66]. Future implementations for autonomous ocean clean up boats include those that can collect trash on a larger scale and support oil spill clean ups.

[70]Ocean Clean Up

Regulatory Implications

Legal Responsibility

Recent autonomous vehicle-related accidents show that legal and regulatory implications are just as important of a consideration as refining the technology itself. Although accidents caused by driver-error will be reduced with the commercialization of self-driving cars, standard regulations are not yet in place to deal with collisions caused by fault of the car. Given the grey area in these cases, especially when the vehicles in question are not yet “fully-autonomous” (ie. Level 4 or 5), there is debate as to whether the human driver or the car itself should be liable for damages.

One interpretation is that pedestrian fatalities at the hands of a self-driving car are still the responsibility of the human operator, even if the vehicle was in fully autonomous mode. However, others argue that in some scenarios, such incidents would have been out of the operator’s control and they should not receive the full blame. This grey area is especially concerning with the idea that at higher levels of autonomy, the vehicles themselves will be granted driving licenses [71], allowing human “drivers” to bypass the requirement of needing to be licensed. While this will bring improved mobility to those who are disabled or visually impaired, it makes knowing who to blame in the case of accidents all the more complicated.

For drivers who are injured by the car, it is also unclear as to whether manufacturers can be held accountable. The US bill, AV Start Act, has been controversial among critics. The act states that a person badly injured while riding in or operating a self-driving car will be unable to take part in a class-action lawsuit or sue the manufacturer [72]. Rather, any dispute will need to be resolved through arbitration, which is said to shift the balance of power in favour of big corporations, who more often than not are the ones hiring the arbitrator. Ridesharing companies, including Uber and Lyft, have already made changes to their terms of service to include forced arbitration.

Responsibility to Passengers

Thus another area of potential regulation is the vehicle’s responsibility to its passengers. The ethicality of decision making by autonomous vehicles is a topic of controversy. When current drivers make the decision to brake their car to avoid hitting a pedestrian, they are making a moral decision. These choices will be ones that self-driving vehicles will also need to make. In the design of the internal algorithms and software that determine these choices, it is difficult to set a universally standard code of ethics, especially when moral principles that guide a driver’s decisions vary by country [73] and from individual to individual.

[73] Worldwide Variations in Moral Principles that Guide a Driver's Decisions

Should the vehicle prioritize the safety of the driver and its passengers? Or take the series of actions that will result in the least number of casualties?

With the differences in moral code among individuals, potential adopters of the technology may be apprehensive about handing over control to the vehicle to make these sort of life-altering decisions. In fact, not everyone is an advocate for fully autonomous or automated vehicles. In recent events, Boeing pilots have expressed discomfort in flying the 737 MAX 8 airplanes. These planes, despite operating at a certain degree of automation, lack the common feature that allows pilots to pull planes out of nosedives and avert crashes in dangerous situations, preventing pilots from being able to take full control at any given time [74]. The Boeing pilots argue that automation should never replace the overriding commandment that the pilot has to fly the aircraft and know what it is doing.

Changes to Manufacturing

In addition to legal regulations, manufacturing regulations will also need to change. Under current US safety rules, a motor vehicle must include traditional controls, such as a steering wheel, mirrors, and foot pedals to be permitted for operations on public [75]. However, with higher levels of autonomy, these traditional controls are no longer necessary as functions will be completely controlled by the vehicle’s internal systems.

Impact on Various Industries

Impact on Insurance Sector

One sector that will be impacted greatly is the insurance industry because the introduction of autonomous vehicles will greatly decrease the number of accidents. As more than 90 percent of accidents are caused by human error, taking the driver out of the equation is going to cause huge changes for insurers. Research from the Stevens Institute of Technology in New Jersey estimated that premiums could drop by 12.5 percent of the total market by 2035 [76]. Research done by Clements and Kockelman from the University of Texas estimated that the size of the insurance industry will drop by $108 billion US which is a 60 percent decrease in industry size [77]. Insurance will undergo a huge change once automation reaches level 4 and 5 making insurance a much less consumer-facing industry. Rodney Parker, an associate professor at Indiana University says this would cause liability to be transferred to the manufacturers and the licensers of the software rather than from the individual [76]. This could result in insurers selling their policies to companies rather than to drivers. Carmakers and the companies that develop the parts would be responsible for any failure of products.

[78] Potential implications for the insurance industry in Canada

Impact on Roads and Parking

Because driving will be out of the hands of the driver and now solely the responsibility of the car, traffic fines, including parking tickets and speeding tickets will be greatly reduced. This results in a large loss of income from traffic fines. A study by Andreas Tschiesner, a senior partner in the Automotive and Advanced Industries sector, found that 30 million euro goes to the city of Hamburg via parking tickets, while Stuttgart receives 11 million euro via speed cameras. This source of revenue will be gone because autonomous vehicles will not speed or park in incorrect places. Because of this decrease, Andreas suggests that a congestion charge gets introduced. For example, the city of London has already introduced a congestion charge within areas of the city with a high volume of congestion [79]. The traffic policing industry in the U.S. is expected to decrease by $5 billion USD which a 50 percent decrease for the industry [77]. A decrease in staffing for traffic police would allow police officers to be more productive in other aspects of law enforcement.

Impact on the Trucking Industry

Another sector that will be impacted heavily by autonomous vehicles is the trucking industry as autonomous driving could replace approximately 294,000 long-distance truck drivers over the next 25 years according to the Wall Street Journal. The reason is because long-distance highway driving is easier than driving through congested city streets. Using autonomous trucks could lower freight costs while improving productivity and safety resulting in a much more profitable industry [80]. The freight transportation industry in the U.S. is expected to increase by $100 billion and an annual percentage increase of 17 percent for the industry [77].

Impact on Jobs

[81] Concept of working while commuting in an AV

Skeptics of autonomous vehicles are afraid that the commercialization of autonomous vehicles will cause citizens to lose their jobs. This is particularly true for driving jobs. A report from Goldman Sachs estimates that the peak saturation of autonomous vehicles could cause U.S. drivers to lose jobs at a rate of 300,000 a year. There was a total of 3.1 million truck drivers in the U.S. which is 2 percent of total employment in America [82]. However, autonomous vehicles will also create many new jobs for the economy. Many companies are currently looking for specialized talent like engineers, technicians, software developers, and designers to build autonomous vehicles [83]. Car maintenance jobs, sales jobs, logistics jobs, and more highly technical or skilled jobs will continue to arise for the economy. The economy-wide effect of autonomous vehicles is estimated to have a productivity boost $448 billion [77]. The reason for this boost in productivity is because autonomous vehicles will free up hundreds of hours per passenger per year so that workers can be productive during their daily commutes [84].

Impact on Medical Industry

Because autonomous vehicles will decrease the number of car accidents, the need for trauma care will fall. More than 2.5 million people visited a U.S. emergency room from car accidents in 2012 [85]. The total cost of these medical costs was $18 billion USD. The decrease in demand for trauma care will allow for the medical industry to focus on population health and preventative care.

In addition to that, patients won’t miss appointments as much as they find new options to reach healthcare providers. Missed appointments at large clinics account for 18 percent to more than 30 percent. With ridesharing, vulnerable populations like the elderly or disabled will have greater access to health care [85]. Not only that, autonomous vehicles can allow health care practitioners to create mobile care units that can be dispatched to remote areas or patient homes.

Also, the mainstream use of autonomous vehicles would lead to shorter commutes and cleaner air for citizens. Cleaner air would help reduce the risk of respiratory diseases, heart disease, stroke, and lung cancer. Also, researchers found that car commuting is associated with poor sleep, higher levels of stress, and exhaustion. Meanwhile, long car commutes are associated with obesity, poor cardiovascular health, high blood pressure, self-reported tension, and negative mood in the evening [85]. Using autonomous vehicles to take that stress off of drivers could then improve health outcomes.

Impact on the Auto Repair Industry

Because of less accidents, there will be a decreased demand for auto repair services. Therefore, mechanics’ skills may not be as necessary with more connected and software-dependent cars. Also, with diagnostics reports of self driving cars, owners will be able to avoid more expensive repairs with preventative maintenance. Also, less accidents means a decreased demand for replacement parts. However, there will be an increase in demand for electronics that will be used for the autonomous vehicle. By 2030, PwC predicts that that electronics will account for 50 percent of automobile manufacturing costs. [86] Overall, the auto repair industry is anticipated to drop by an estimated $15 billion USD which is a 26 percent decrease in the auto repair industry size [77].

[77] Economic impact of Autonomous Vehicles on various sectors in the U.S.

Future Direction

Thus far we have talked about the exciting possibilities of autonomous vehicles and what they could bring to consumers and societies. We will now look at potential impacts of mass-adopted AV's as well as how we may plan for the future.

Perception and Preparation

Public perception of AV's

[87] American Attitude Towards AV's

The public is increasingly aware of the technology and possibilities AV’s could bring to our lives, but for actual adoption to take place, there needs to be a positive sentiment towards the technology and a willingness to adopt it. Most studies we found were based in the US, though we believe it is reasonable to extend this sentiment to encapsulate Canadians in urban cities as well. One study was summarized by the following 3 points [88]:

- Public perceptions play a crucial role in wider adoption of autonomous vehicles (AVs).
- As the public increasingly interacts with AVs, their attitudes toward the technology are more likely to be positive.
- Policy-makers should pass legislation that will authorize testing and operation of AV's on public roads.

According to a study conducted in 2017 by PEW Research, it would seem the public is still leaning towards uncertainty towards AV’s in their everyday lives. There is a lot of buzz, but it is not common to see or experience the cars, therefore the public is cautious, and rightly so, about all the benefits AV’s chalk up to be.

However, many people are expecting AV’s to be amongst us in the next 10-50 years [87]. Therefore, it seems to be this looming topic that is supposedly going to change the way we live, but until we start actually seeing the cars on roads around us for testing or education on the topic ramping up, the study suggests the public will remain uncertain about AV’s. We thus enter into a type of chicken and egg paradox - will cities take the lead, potentially offering city wide AV’s to warm the public to the idea (e.g. more widespread Waymo car as mentioned earlier), and the public will follow, or will the public take the lead, and cities have to play catch-up? If history repeats itself, then we will most likely fall under the latter - similar to the invention of the car in 1920, the public adopted it en-masse and cities, regulators and the likes played catch-up.

How are cities planning for the future of cars?

As it currently stands, us consumers are waiting on the sideline. Although Tesla is leading the charge with it’s semi-autonomous vehicles, plans to release driverless cars are non-existent at the moment. We mentioned earlier what companies are doing to make and develop their technologies to create driverless cars, however without the cities taking the proper steps to make cities adaptable or make the public feel safe, it is doubtful we will see adoption of AV's. Therefore, are cities taking this imminent possibility seriously and if so, what steps are they taking to embrace the change? Typically, as it is with most modern changes, it is the bigger cities that are ahead. Below are a few examples of the steps cities are taking to help accommodate AV's. Most cities, at this stage are working with companies as intermediaries to facilitate and (no pun intended) pave the way to new regulations and laws that will support the technology. Additionally, for initial testing, cities are mostly experimenting with municipal vehicles, as this is safer and easier to control for the early stages.

RPA's 2040 vision for city adoption [89]

Use cases

Arlington, TX

In August 2017, Arlington began testing a 12 passenger AV shuttle on a fixed route on non-public roads. There are currently two shuttles in use, and instead of drivers on board, there are “information ambassadors” that are helping educate the passengers and answer any questions they may have [90].

In October 2018, Arlington became the first Texas city to offer an on-street driverless transportation option to the general public with partner The pilot program came to an end recently, in May 2019 and was deemed a success, logging 440 miles and 760 trips over the two years. There was positive sentiment from the riders with 98% saying they felt safe, and 91% saying they thought their ride was smooth [91].

Chandler, AZ

Chandler Arizona is a suburb outside of sunny Phoenix home to many tech companies. General Motors’ IT Innovation Centre is situated in this hub, and they have partnered up with the city on a few projects. AV’s have been tested in this city since 2015, and on public roads since 2016. The city is filling the helpful role of being a facilitator for AV’s, with the state as the regulator. The city has tested emergency vehicles on a closed track, and has amended its zoning development code to reduce parking requirements up to 40%. The latter is huge, as most US cities have traditionally been built around the modern day car, with zoning codes requiring houses to be far apart and demanding a certain percentage of the city to be dedicated to parking spaces [92]. Now, in Chandler, after the change in code this has freed up space for passenger loading zones for drop-off and pick-up spots for AV’s [90]. If you are lucky, you can sign up to be a rider part of Waymo's rider program (400 at the moment), and get driven around in pilot vehicles [93]. Chandler is continuing to prepare for autonomous vehicles, the Waymo ride-share program opened up to public testers in March 2019, and new buildings with carparks planned, are being thought over and asked to be "ready for conversion" for alternative uses, should driverless cars be mass adopted. With their current programs in place, Chandler has seen 21 incidents related to their driverless cars, which is very low given their city population of 250,000.

Vancouver, BC

[94]Vancouver's 2040 Transportation Plan
In August 2018, Vancouver received $386,000 in funding from the federal government to study the future of AV’s, part of Transport Canada’s program to advance connectivity and automation in the transportation system. Similar to Chandler, the city will initially investigate how AV’s may be introduced for municipal vehicles such as police cars, firetrucks, etc. There is no timeline (at the moment) or a plan for when driverless cars will be on the road in Vancouver. As mentioned earlier in our wiki, one essential part of a driverless world is connectivity – something Vancouver currently lacks [95]. Therefore, there is recognition of the shortfalls, but no current spending plan to revamp such infrastructure. With 5G being scaled out, this should change in the coming years. In the mean time, Vancouver will continue to research AV’s and develop its 2040 Transportation Plan.

Toronto, ON

Similar to Vancouver, Toronto also received $370,000 in funding from the federal government to study AV's in their city. With the funds, Toronto has been able to further some existing projects and start some new ones. They have a great section on their city's website that shows a complete timeline of everything they've done thus far: Like most cities, Toronto's focus for the moment is on municipal vehicles, in this case they decided to focus on using AV's to "close the last mile" in their public transportation offerings. As it stands, there is a "low-demand" gap that is difficult to fill, as it does not necessarily merit a full fledged bus or street car stop, however there is still demand for public transportation. The city of Toronto released an RFI that closed January 2019, which was to finalize the technology and neighbourhood to place the vehicle. The aim is to have a "highly automated" vehicle that will carry 8-12 passengers at a time with operators on board during the pilot phase [96]. The trial is scheduled to start in mid 2020 and last 6-12 months (future groups, if you are reading this, please update on this development!). This is a great example of a city using AV's to augment current city offerings where traditional transportation would not make as much sense, as well as exposing the public to new technologies and warming them up to the idea of driverless vehicles.

Where to Invest Remains a Debate

The unparalleled efficiency of busses compared to individual cars
Although currently most cities are testing or using AV's for municipal related vehicles, there is significant worry or debate about the future of public transportation within cities and whether it will become redundant or not. If we can make busses independent, that's a great step. However, if we can get individual cars to be independent, and low cost, there could be significant less users of public transportation, as people generally prefer the comfort and privacy of individual cars [97]. If everyone was using driverless cars, even if they are shared, we would see an increase in traffic and we might be worse off, so would we go back to more efficient modes of transportation? As much as there are cities that see AV's as tools to bolster public transportation systems (e.g. Toronto above), on the other side of the picture, there are cities that are killing transportation bills and reducing city funding of public transportation in favour of investing in driverless, personal vehicles. The New York Times [98] wrote a great piece about the debate cities are faced with is whether to take a gamble, bet that driverless cars will be adopted en-masse in the near future, and thus focus investments there (at cost of public transportation) or to keep as is [99].

However keeping things as is does not free up the capital needed for moving driverless cars forward. That is where private companies have come into play, as we've pointed out in our wiki here they are advancing the technology tremendously. Cities are going to be forced soon enough with decisions related to infrastructure - to build a parking lot, or to build more charging stations, to build more subway stops, or repave underground tunnels to accommodate more cars. Ultimately these decisions will be shaped by the public, regulators and lobbyists. Cities today should continue to expose and introduce the concept of AV's and over time, the slow adoption will be rolled out hopefully benefiting us all.


In our wiki we touched on various areas under the broad topic and ever changing autonomous transportation industry. In our historical overview, we recollected all major developments that lead to where we are today in terms of technology of autonomous land vehicles, as well as what companies are currently doing in the market. Then we discussed the technical details and technology behind autonomous cars and the various stages we still need to get through. The goal is to eventually get to level 5 automation, where the car is completely automated and there is no need for a driver. This may seem like a farfetched idea right now, though we are getting closer and closer. We then dived into two other industries that are seeing developments on the automation front, first the trucking industry and then the boating industry. Both are heavily involved in worldwide logistics, moving goods and people around the world and could thus benefit tremendously from advancements in automation. However, there are some downfalls to all these developments, which we discussed in our future looking section, where we started with impacts on various industries. Well established and mature industries such as insurance or legal fee systems will be challenged and need to be revised with the introduction of autonomous vehicles. Simply put, we have implicitly built the notion of a human driving our vehicles into these areas, thus when we remove this, problems arise. Additionally, jobs could be at stake and have major effects on countries and economies. Finally, we brought the topic back to the autonomous vehicles every day consumers may soon see, as this is the area most relevant to our everyday lives. We discussed what measures cities are taking to embrace or push back against AV's, as well as the general perception from the public to such developments. The potential for autonomous vehicles is massive, and should it deliver on its promises we could radically change the way we live. There are obstacles ahead, and more than anything, we should be excited and embrace these coming changes, while still being vigilant in terms of safety, regulations and how AV's might impact our societies.


Chester Mew Jean-Paul Morneau Trevor Quon Kady Wu Cynthia Zhao


  4. 4.0 4.1
  7. 7.0 7.1
  8. 8.0 8.1
  15. 15.0 15.1
  21. 21.0 21.1
  22. 22.0 22.1
  25. 25.0 25.1
  30. 30.0 30.1
  37. 37.0 37.1
  42. 42.0 42.1
  54. 54.0 54.1
  59. 59.0 59.1
  62. 62.0 62.1
  66. 66.0 66.1 66.2
  69. 69.0 69.1
  73. 73.0 73.1
  76. 76.0 76.1
  77. 77.0 77.1 77.2 77.3 77.4 77.5
  85. 85.0 85.1 85.2
  87. 87.0 87.1
  90. 90.0 90.1
Personal tools