TU-Automotive Detroit 2017: Accommodating the autonomous vehicle

Influences of the autonomous car on technology, society, and user experience

2017/07/06

Summary

 The TU-Automotive Detroit 2017 exhibition and conference was held from June 7 through June 8, 2017, in Novi, Michigan, U.S. The conference had over 3,000 attendees, featured approximately 150 speakers and showcased displays from more than 130 exhibitors. Though autonomous technology was still a major theme, the conference highlighted other topics including business strategy, urban mobility and user experience. The overarching theme of connectivity was prominent during the conference, as many sessions noted the connected car’s significance in the future.

TU-Automotive Detroit 2017 exhibition floor
TU-Automotive Detroit 2017 exhibition floor

 This report is the third of three reports which focuses on the discussion sessions presented at TU-Automotive Detroit. This report focuses on presentations related to autonomous technologies, mobility and user experience within the vehicle. Two reports, one focusing on business models and strategies, and another on connectivity considerations and technologies, were previously released.

Related reports: TU-Automotive Detroit 2017
Business Strategies and Opportunities in Changing Ecosystem
Connectivity Perspectives, Challenges and Technologies



External outlook of autonomous vehicle technology

Drawing experience from data-driven industries

Session: Collaborating in the Race to Automation
Company Speaker Title
Intel Doug Davis Senior Vice President & General Manager,
Automated Driving Group

 Throughout his presentation, Doug Davis emphasized the importance of developing and implementing autonomous technology faster than other vehicle safety technology. In the past, automotive safety innovations have traditionally required a significant amount of time from initial development to being mandated. For example, airbags were patented in 1951 but were not mandated in the U.S. until 1998. ABS was invented in the 1950s and was not mandated until 2013, while three-point seat belts were developed in 1959 and mandated in 1968. The potential benefits of autonomous vehicles encourage a quicker implementation method, as Doug Davis noted a figure from NHTSA which stated that approximately 93% of fatal accidents are caused by human error. Beyond reducing the number of accidents, autonomous driving also provides a significant economic benefit. A study from Morgan Stanley said that the autonomous vehicle has the potential to generate USD 1.3 trillion in savings for the U.S. economy.

Speakers of "Fireside Chat with Ford Smart Mobility"
session
Economic and societal benefits of autonomous vehicle
Source: Morgan Stanley

 Studies from IHS forecasted that total vehicle production in 2030 will be between 105 million and 125 million units, from the 94 million units produced in 2016. However, the computational power necessary in the car is expected to increase at least tenfold from 0.5-10 TFLOPs (floating point operations per second) currently to 50-100 TFLOPs by 2030. In addition, the number of pixels and amount of storage necessary for a vehicle is expected to increase by 1,000 times while using less ECUs. Doug Davis noted that while the data currently available is predominantly human-generated, data in the future will predominantly consist of machine-generated data.

Speakers of "Fireside Chat with Porsche Digital"
session
Shift in computational utilization in the future vehicle
Source: Intel

  Considering the transformation of the automotive industry in data usage, Doug Davis gave examples of other industries that have experienced similar changes. In the PC industry, the first personal computer was available in 1975, with the market featuring more than 100 incompatible brands by 1980. In 1981, IBM launched the first PC with standard off-the-shelf hardware, leading to two million PCs being sold in 1984. The server industry was similar, shifting from mainframes to microcomputers in the early 1990s. By the mid-1990s, servers transformed into enterprise server rooms through standardization and have currently evolved into cloud service providers. These examples illustrate the importance of both standardization at the base level and collaboration with other companies, in that industries are able to accelerate implementation and quickly scale operations once standards are in place.

 

Autonomous vehicles impact on the state

Session: Autonomy: The State Perspective
Company Speaker Title
Michigan Department of Transportation (MDoT) Kirk Steudle Director
Nevada Department of Motor Vehicles April Sanborn DMV Services Manager III – Driver Programs
Wall Street Journal Moderator: John Stoll Chief Reporter

 The panel session, “Autonomy: The State Perspective” began with both panelists providing a brief overview of their state's recent activities relating to autonomous vehicles. Kirk Steudle mentioned legislation from December 2016 enabling the legal operation of self-driving cars beyond testing on Michigan public roads. Michigan also enacted legislation which created a council targeted towards future mobility. The council makes policy recommendations in areas including automated and driverless vehicles. April Sanborn discussed legislation to broaden testing in Nevada, and allow for a highly restrictive autonomous vehicle driver’s license for quadriplegics.

Speakers of "Fireside Chat with Porsche Digital"
session
Panelists of “Autonomy: The State Perspective” session. From left to right: April Sanborn, Nevada Department of Motor Vehicles; Kirk Steudle, Michigan Department of Transportation; Moderator – John Stoll, Wall Street Journal

 In discussing the role of the state, both panelists agreed that states have specific roles and that “states do not belong inside the car”. Regulations and mandates regarding the design and development of autonomous vehicles are issued by NHTSA, while state governments focus on operation and insurance. Autonomous vehicle testing, while more ambiguous, is also currently regulated by state governments. Both speakers noted that state legislation will give precedent to federal legislation for autonomous technology. Despite the specific governance, states have been proactive in developing regulations for autonomous vehicles. Both Nevada and Michigan have been open to meet with autonomous technology manufacturers for discussions and have allowed companies some leeway in regulation. For example, April Sanborn discussed an instance where a company developing a low-speed autonomous vehicle had trouble meeting the 10,000-mile testing requirement. The Nevada government has been in discussions with the company to bypass this requirement.

 The panel session also provided insight on how states are reacting to possible changes caused by the autonomous vehicle. For example, when looking at the changes in revenue for the state, April Sanborn noted that revenues from vehicle registration could possibly decrease. Kirk Steudle noted the possibility that driver’s licenses would be less common. Thus, new government revenue streams would be needed. Another issue was the potential job loss in fields such as trucking. Both speakers made analogies comparing the autonomous vehicle to the affordable car which replaced the buggy and horse, and said that businesses would have to change based on the demand from consumers. Kirk Steudle also discussed how the autonomous vehicle might change traffic laws and infrastructure. He noted that possible changes to traffic laws include relaxing regulations on following distances and speed limits. In terms of infrastructure, Kirk Steudle believed that it was too early for standards in infrastructure, though there was discussion on potential implementation.

 

Encouraging acceptance of autonomous technology

Session: Consumer Acceptance of Automation
Company Speaker Title
National Safety Council Alex Epstein Senior Director, Digital Strategy & Content
Mitsubishi Electric Gareth Williams Director, Advanced Development
J.D. Power Kristen Kolodge Executive Director, Driver Interaction and HMI
Tourmaline Labs Lukas Kuhn CTO
AAA Moderator: Greg Brannon Director, Automotive Engineering

 During the panel session, “Consumer Acceptance of Automation”, panelists discussed the difficulties facing consumers with the increasing prevalence of autonomous technology and advanced driver assistance systems (ADAS). One issue involved the variance and overall lack of standardization for various ADAS functionalities and names. The panelists agreed that some standardization would benefit the consumer, as Kristen Kolodge described the current aggregate of ADAS technologies as an “alphabet soup” with varying names, acronyms and functionalities. For example, Alex Epstein mentioned the case of one company's emergency braking system including pedestrian detection, while another didn't. He presented the idea of automakers forming a brand-neutral campaign between themselves emphasizing best practices and common themes to develop a form of standardization to accommodate customers. However, he also cautioned against overpromising and overstating the capabilities of certain autonomous technology. Lukas Kuhn suggested a different viewpoint, stating that the technologies should be similar to the iPhone and be more intuitive and natural, while also providing visible value to the customer.

Speakers of "Fireside Chat with Porsche Digital"
session
Panelists of “Consumer Acceptance of Automation” session. From left to right: Alex Epstein, National Safety Council; Gareth Williams, Mitsubishi Electric; Kristen Kolodge, J.D. Power; Lukas Kuhn, Tourmaline Labs; Moderator – Greg Brannon

 Another issue discussed in the panel was the amount of knowledge that a customer should have to utilize autonomous technology. Gareth Williams said that ideally, users should understand and be comfortable with the system’s capabilities, and also know how to engage and disengage the system. For example, displaying additional information on the instrument cluster for a pedestrian-detecting emergency braking system would help the user understand the situation. Lukas Kuhn provided a counterpoint however, that many customers do not understand current automotive systems , and the more important issue is for users to understand how to interface with ADAS technologies. Kristen Kolodge suggested that the user needs to understand the role of the driver and how the driver interacts with the systems

 Finally, the panel discussed the idea of reluctance to use ADAS and autonomous technologies due to the loss of control and psychological factors. Alex Epstein noted that cars have historically been marketed as a symbol for freedom. Thus, when certain aspects of the vehicle are no longer controlled by the driver, the generally established perception of the vehicle is broken, creating a negative feeling. He also gave an analogy comparing the stage of autonomous technology development to a child growing into an adult in that autonomous systems need more development and improvement. Lukas Kuhn mentioned that society is in the middle of a generational shift, in which the younger generation is less concerned with being able to drive and control the vehicle and more amenable to the use of self-driving cars. One method to improve consumer acceptance, proposed by Kristen Kolodge, is to increase experience and exposure with autonomous features and technologies over a period of time.

 



Technologies to support autonomous development

New radar technologies for autonomous requirements

Session: Kiss SiGe goodbye: Making radar smaller and more precise
Company Speaker Title
Texas Instruments Sudipto Bose Marketing Director

 The presentation, “Kiss SiGe goodbye: Making radar smaller and more precise” focused on the sensing capabilities required by autonomous vehicles and showcased Texas Instruments’ recently developed radar solution designed to meet those demands. As vehicles include more autonomous and ADAS features, Sudipto Bose stated that the sensing requirements needed in the vehicle will increase as well. He summarized the sensing capabilities which a sensor needs into three requirements.

  1. The ability to detect range, velocity and angle of objects and to classify an object into categories such as pedestrian, bicyclist, or motorized vehicle
  2. The ability to penetrate plastic, glass and other materials without being too large
  3. The ability to reliably function in adverse environmental conditions such as rain, fog, dust, bright light and darkness

 Currently, the only sensing technology that satisfies all three conditions is radar.


 Though radar is capable of achieving all three conditions, most radar systems lack the level of precision needed for more sophisticated autonomous and ADAS functions. Sudipto Bose provided a list of requirements including resolution accuracy levels of less than five centimeters, angle accuracy levels less than one degree, multi-modal operation and object classification. In May 2017, Texas Instruments launched its new millimeter wave single-chip Complementary Metal-Oxide Semiconductor (CMOS) sensor family, capable of achieving the requirements previously mentioned. The CMOS sensors are 10.4 mm x 10.4 mm, have a range resolution of less than four centimeters and range accuracy of less than 50 micrometers. Compared to standard sensors, the millimeter wave sensors reduce both power consumption and board space by 50%. The sensors also operate accurately through various surface mediums and in various environmental conditions. Potential automotive applications utilizing these sensors include frontal long range radars, multi-modal radars, and short range radars.

Panelists of "Unlocking New Automotive Revenue Streams"
session
Description of gap of current radar technology and future radar needs
Source: Texas Instruments

 

Role of mapping technologies in autonomous vehicles

Session: Mapping an Autonomous World
Company Speaker Title
TomTom North America Paul Hohos Vice President Sales, Automotive

 Paul Hohos’ presentation, “Mapping an Autonomous World” focused on the role of mapping in future autonomous vehicles and the development of certain mapping technologies. An accurate map system gives the autonomous car the ability to see areas beyond its sensors. This provides benefits such as navigation, localization in knowing the location of the vehicle, perception in adding data redundancy, and path planning. However, for autonomous vehicles to use maps effectively, the maps themselves need to be regularly updated to avoid providing outdated data. TomTom features maps that are developed through a combination of traditional methods such as mobile mapping and field surveys, and community input, which obtains data from vehicle sensors and probes. Much the data provided in TomTom’s intelligent mapmaking process is based on live GPS traces of over 550 million GPS devices globally.

 As vehicles increase in capability and include more advanced technologies, the maps provided must also be able to provide more detailed information. Paul Hohos described a series of map layers which could be utilized by a vehicle, with each layer containing a separate set of information. At the base layer, the map contains information about the road network and directionality. Higher layers contain information about addresses, traffic signs, points of interest, 3D visualizations and ADAS information.

Panelists of "Unlocking New Automotive Revenue Streams"
session
List of map layers and applications
Source: TomTom

 Paul Hohos showed three TomTom advanced map technologies. HD Map provides an accurate and realistic roadway representation, and includes information on lane geometry, lane markings, center lines, lane dividers and road boundaries. Additional attributable information available on HD Map includes speed restrictions and lane widths. RoadDNA is a tool that provides an accurate optimized 3D lateral and longitudinal view of the road. By combining this data with sensor data, the vehicle can determine its precise location. RoadDNA has relative accuracy less than 15 centimeters laterally and less than 50 centimeters longitudinally. TomTom has also developed an artificial intelligence (AI) to help identify traffic sign information. The locations of these signs are recorded by a survey vehicle’s radar, and are then processed by the AI through a database to classify and provide information about any specific sign.

Comparison of MQB platform
battery and MEB platform battery
Overview and sample visualization of HD Map
Source: TomTom
Volkswagen I.D. Buzz
Concept
Overview and sample visualization of RoadDNA
Source: TomTom

 

Benefits of V2X and ADAS fusion

Session: V2X and Autonomous Driving is a “two-way street”
Company Speaker Title
u-blox Costas Meimetis Director, Product Strategy

 The presentation, “V2X and Autonomous Driving is a ‘two-way street’”, given by Costas Meimetis, highlighted several use cases and benefits of the fusion between V2X and ADAS technologies. One example currently available is the use of cooperative adaptive cruise control, or platooning. With both V2X and ADAS, vehicles can not only track the following distance of the vehicle immediately ahead but also the speed of other vehicles in the platoon. This minimizes jerking movements and thus, optimizes traffic and fuel efficiency. Costas Meimetis noted that there are more than 250 million vehicles currently operating in the U.S., making it difficult to see high levels of technology penetration quickly even with an aggressive launch. He then provided another application, noting the benefits of improving overall awareness even with a low V2X penetration rate, by showing one car communicating with another to let it know when to perform an overtaking maneuver. A graph was shown which emphasizes how the fusion of V2X and ADAS improves environment awareness even at lower penetration rates.

Comparison of MQB platform
battery and MEB platform battery
Example of ADAS and V2X fusion for overtaking maneuver
Source: u-blox
Volkswagen I.D. Buzz
Concept
Graph showing environment awareness based on technology penetration rate for various systems
Source: Gunther, Trauer, Wolf

 Other examples of V2X and ADAS fusion include the sharing of information outside a vehicle’s sensor range, such as a pedestrian walking across a street that is obscured by another car, or cooperative collision avoidance, where vehicles communicate to reduce possible damage in an emergency. One prerequisite in all of the use cases, is the importance of accurate positioning in V2X applications. Current standards state that global navigation satellite systems (GNSS) need to have an accuracy of 1.5 meters. However, in the future, accuracy will need to be improved down to the decimeter level for usage in ADAS and V2X applications. Costas Meimetis showcased the improvements made in the u-blox automotive dead reckoning (ADR) 4.10 system, which calculates positions in three dimensions based on motion sensors and speed data when satellite signals are blocked in order to determine a vehicle's current position.

Panelists of "Unlocking New Automotive Revenue Streams"
session
Visualization of improvements in location accuracy demonstrated by u-blox ADR 4.10 system
Source: u-blox

 



Autonomous technology driving societal and interface changes

Future changes in society due to the autonomous, connected car

Session: Changing the Shape of Society with Autonomous Vehicles
Company Speaker Title
Harman International Steve Surhigh Vice President & General Manager,
Applications Services

 Before describing the changes expected in society with the development of the autonomous vehicle, Steve Surhigh explained how society changed with what he described as Mobility 1.0, the launch of the affordable automobile. There were five aspects of society that changed with Mobility 1.0: infrastructure, employment, lifestyle, housing, and business models. The changes in society that will occur with the development of the autonomous and connected vehicle, described as Mobility 2.0+, can also be organized into the five aforementioned categories. However, the new changes that will occur are notable in that it will have both physical and digital effects across most of the societal categories, as shown on the table below.

Category Changes due to automobile Changes due to autonomous automobile
Infrastructure -Dirt roads developing into interstates
-Urban communities expanding to suburban communities
-Development of street lights, traffic signals, road signs
Physical: -Reduction of parking lots
-Increased amount of green space
-Elimination of signs, traffic signals, etc.
Digital: -Addition of sensors and related infrastructure
-Increase in computing power and storage
-Development of systems to monitor road conditions and traffic
Employment -Development of higher paying jobs
-Creation of jobs outside of cities
-Increased demand for industrial supplies
-Decreased employment as taxi drivers, bus drivers, traffic officers and truck drivers
-Increased employment as hardware and software engineers, data scientists, car mechanics, logistics and fleet management
Lifestyle -Enabling freedom to travel
-Enabling concept of family vacations
-Improving overall air quality
Physical: -Increased free time
-Improved safety
-Improved mobility for children and elderly
Digital: -Increased access to real-time information
-Increased development of applications enabling mobility
-More customized and individual tailored experiences
Housing -Development of suburban communities
-Movement of population away from rail lines and waterways
Physical: -Elimination of garages and driveways
-Location and distance becomes less relevant
Digital: -House exchanges data with various IoT devices
-Systems within the house capable of anticipating needs
Business Models -Creation of gas/service stations
-Creation of fuel taxes
-Enabling more effective shipping of products
Physical: -Enhanced effectiveness of delivery services
-Change in insurance and legal restrictions
-Vehicle ownership changing into a business opportunity
Digital: -Bandwidth providers to have new in-vehicle opportunities
-Monitoring and managing vehicle fleets
-Monetizing in-vehicle experiences

 Steve Surhigh described three core areas in autonomous vehicle development: vehicle operations, in-vehicle content, and vehicle communication with various ecosystems. These three areas serve as the primary drivers enabling the societal changes described above. Furthermore, in preparation for these changes, OEMs need to pursue certain activities, such as the development of connectivity technology to 5G and beyond, collaborate with other companies and industries to make data more accessible, enhance machine learning capabilities, develop contextual virtual assistants with augmented reality and intelligence augmentation to create unique experiences, and improve big data management.

 

Redesigning HMI for future considerations

Session: HMI: Re-Imagining In-Car Experiences
Company Speaker Title
Smart Design Edward Laganis Design Engineering Director
SRI International Ajay Divakaran Technical Director, Vision and Learning at Center for Vision Technologies
Panasonic Michael Tschirhart Manager, Holistic Innovation, Advanced Engineering
Mitsubishi Electric Jacek Spiewla User Experience Manager, Advanced Development
Mapbox Alex Barth Vice President, Business Development
WardsAuto Moderator: Drew Winter Editor in Chief

 The panel session, “HMI: Re-Imagining In-Car Experiences” discussed a variety of topics about how the in-car experience would change with the addition of more advanced technologies. One topic of discussion was the role of voice interaction in the future. Ajay Divakaran noted that, voice commands currently lack context and are often one-time interactions. He suggested that voice commands be incorporated in a more holistic approach, for example with driver monitoring and emotion recognition to give the system more personalization. Jacek Spiewla pointed out that voice commands currently exist as a redundancy, mainly for the convenience of usage while driving. As driving may not be the primary task in future vehicles, the HMI needs to be adjusted to fit the paradigm shift.

Panelists of "Unlocking New Automotive Revenue Streams"
session
Panelists of “HMI: Re-Imagining In-Car Experiences” session. From left to right: Edward Laganis, Smart Design; Ajay Divakaran, SRI International; Michael Tschirhart, Panasonic; Jacek Spiewla, Mitsubishi Electric; Alex Barth, Mapbox; Moderator – Drew Winter, WardsAuto

 Another topic discussed was the idea of developing an iconic HMI design that is not distracting. Tesla’s center screen was brought up as an example, which Edward Laganis noted that he had mixed feelings towards, mentioning both ergonomic issues and lack of tactility. He said that there were benefits to using more tactile controls like buttons and felt that it would be better to utilize a platform adapted to the vehicle’s purpose instead of a single display. Michael Tschirhart noted that current technology enables much broader capabilities with systems that gather information about the driver and the environment. This information can then be used to generate experiences specifically tailored to the user.

 In regards to the idea of keeping technology current and making the comparison to the smartphone, Alex Barth pointed out that the car experience differs from the phone experience and should not try to emulate the phone. Instead, developers should look to find ways to evolve the user experience as usage of the system continues. When describing needs for future HMI systems, Michael Tschirhart mentioned the importance of increasing collaboration with other parties to develop solutions across the entire journey. Alex Barth focused on applications with a simpler, easier-to-use and cleaner interface, noting that increasing ease and accessibility will improve the user’s experience. Focusing more on the changes created by the autonomous vehicle, Edward Laganis said that systems will need to be more passenger-centric as the driver’s role will be reduced or eliminated.

------------------
Keywords
TU-Automotive, Autonomous, ADAS, HMI, Mapping, Radar, V2X

<Automotive Industry Portal MarkLines>