The 25th Vehicle Displays and Interfaces Symposium was held September 25-26, 2018 at the Burton Manor Conference Center in Livonia, Michigan. The conference, organized by the Metro-Detroit Chapter of the Society for Information Display (SID) and managed by Palisades Convention Management (PCM), grew to a record 742 registrants this year from 558 in 2017.
The conference opened with a keynote address by Kristin Kolodge, Executive Director of Driver Interaction and Human machine Interface for J.D. Power in Troy, Michigan. Her topic: “Voice of the Customer: Displays as a Trust Enabler.”
Kolodge’s major point was that consumer/driver/passenger acceptance of advanced driver assistant systems (ADAS) and full automation depends on trust. So it’s essential to understand the nature of trust and to learn how to foster it. In this new era, “the definition of quality must expand to incorporate trust. If the consumer can’t understand a feature, it’s a quality problem even it is working as the engineer intended.”
“ADAS is the stepping stone to autonomous vehicles,” Kolodge said, and the stepping stone is solid. Consumers like ADAS and would generally order the systems on their next vehicle. Utilization of the features is about 100%, with the main recurring complaint being “some negative experiences with lane keeping.”
“Take rates” for ADAS are increasing. From the 2017 to 2018 model year, the take rate for forward collision warning and avoidance systems increased 17%; for lane-keeping and centering systems, 15%; for adaptive cruise control, 13%, and for blind spot monitoring and warning systems, 11%. Many members of the audience seemed surprised to hear that 56% of people reported that ADAS prevented an accident in the first three months of use.
In short, the systems are technically successful for the most part and buyers appreciate them, once they understand them. And there’s the rub. Buyers report that less than half of dealers explain the ADAS features, which is all the more damaging because of the lack of standardization in displays, acronyms and symbols.
Larger displays would allow designers to spell out a feature instead of using an acronym and symbol, and this would be a major contribution.
When drivers were asked where they would like technological assistance with the task of driving, “seven of the top 10 [requested] technologies help the driver to ‘see.’ ” These technologies are smart headlights, camera rear-view mirror, advanced windshield display, electronic window tint adjustment, lane-change assist, and emergency braking and steering.
“Trust is critical to building interest in automation, and inherent to starting the conversation. Displays can make a critical contribution to enhancing trust and, therefore, quality.”
The first question in the Q&A session, from James Goel, Director of Technical Standards for Qualcomm Canada, reversed the point of view of Kolodge’s talk. Goel observed that autonomous cars drive themselves differently than human drivers drive. For instance, they come to a full stop at stop lights. Most accidents involving autonomous vehicles, said Goel, occur when human-driven cars strike the autonomous vehicle. Autonomous vehicles could become “uncool” because other drivers don’t trust them. Kolodge answered that this is a crucial issue, and the Mcity data should help understand the issues. (As part of its Mcity project, the University of Michigan is instituting a level 4 autonomous campus shuttle service, which will provide data on autonomous vehicle operations.)
Q: Bob Donofrio. Are insurance companies lowering rates for ADAS and autonomous vehicles
A. Insurance companies anticipate fewer accidents but higher cost of repairs for each accident. Consumers anticipate lower premiums.
In his invited address, “Evolving Technologies and Supply Chain in Automotive Display Segments,” Brian Rhodes of IHS Markit reported his company’s finding that use of analog instrument clusters is declining by 10% per year “as the dash goes digital.”
Rhodes noted that use of combiner HUDs will grow in the near term but windshield HUDs will dominate in the long run. Rhodes said that “AR-ready” HUDs will appear by end of this year, but really just big HUDs that are “AR-ready” only in the sense that they have the large field of view AR requires. Real AR HUDS will appear starting 2020, but only in large, high-end vehicles. The “large” here is important because a projection system capable of producing a large field of view can only fit behind a commodious dash.
Rhodes also noted that “button and knob” makers are buying into the display business, so we can expect icreasing market fragmentation.
Jennifer Colgrove of Touch Display Research predicted that µLED displays will appear in cars during the 2022-23 period. Considering the automotive industry’s three year development cycle and that there are no µLED displays on the market now, that projection may be optimistic. Colgrove said that miniLEDs — that is, miniLED backlights used with LCDs — will appear earlier. That’s a safe bet.
In “Global Display Industry Markets and Technologies,” Bob O’Brien of Display Supply Chain Consultants (Ann Arbor, Michigan) observed that since the auto industry is a relatively small market for displays, it doesn’t drive the display industry as it does suppliers of other components.
O’Brien identified the ten largest makers of automotive displays. In order, they are Continental, Visteon, Panasonic, Denso, Bosch, Nippon Seiki, Alps Electric, Magneti Marelli, Aptiv, and LG. He noted that LG Display has lower margins than the industry average because “OLED displays have been loss-making.” He also observed that BOE will be the biggest display producer, and that China Star (CSOT) is also rising.
Samsung has refused to jump on the OLED-TV bandwagon (well it tried! Man. Ed.), and has directed its attention to electro-emissive quantum-dot (otherwise known as “true QLED”) technology. Now, said O’Brien, Samsung is also looking at QDOLED. This technology uses blue OLED and quantum dots to convert some of the blue light to red and green. Since the blue emissive OLED material is one of OLED-TVs major problems, pursuing this idea seems like a very strange thing to do. Using a blue inorganic LED as the light source would make more sense. Oh. We’re already doing that, aren’t we?
The featured evening speaker was Nadine Sarter of the University of Michigan. (The conference organizers believe in long days but, in their defense, there is absolutely nothing else to do in Livonia.) Her presentation was “Design of Automation Behavior and Interfaces in Support of Human-Machine Teaming.” Sarter took the position that the most effective way to develop increasingly automated systems is to think of them as human-machine teams.
A lot of her work is on the failure of complex systems, and these failures are often due to the machine or the human (or both) not knowing what the other expects it to do. Said Sarter, “You don’t necessariy want more trust; you want trust calibrated to what is appropriate.”
Some interesting ideas came out in the Q&A. It is necessary for the human to have a mental model of the system. That means that newer vehicles will require more training. Starter agreed and said she is a licensed pilot. Sse went through four weeks of daily training to qualify on an Airbus A320, and this was mostly to learn the systems.
Sarter said she will not touch a phone when she drives. She does the studies and analyzes the data. The extent to which phones interfere with driving depends on contents of call. If the content has spatial content (directions to a particular location) it is “horrible.”
Yuge Huang, a Ph.D. student of Shin-Tson Wu at the University of Central Florida, received a 2018 SID Metro Detroit Academic Award and presented the paper “MiniLED Enhanced High Dynamic Range LCD for Automotive Displays.” The main point of miniLED backlights is to do local-area dimming with many zones in a very thin and lightweight package.
In one part of her paper, Huang noted that a 12.3″ cluster display with intrinsic contrast of 2000:1 requires about 950 dimming zones to eliminate the halo effect. In subsequent communication, she said it would take about 364 zones to make the halo effect invisible on a 65″ TV viewed from six feet (2m). This informed a spirited conversation that was growing out the Value Electronics TV Shoot-out, that had been held two days before. In that shoot-out, a 65″ Samsung Q9F with 480 zones showed a serious halo effect. My own subsequent observation of a Vizio P9 Quantum with 192 zones showed halo that was visible but much less so than the Q9F.
It seems that halo is a surprisingly complicated animal. I look forward to Huang’s next paper on the subject. (KW)
This summer, as a part of its Mcity initiative, the University of Michigan began an automated shuttle service on its North Campus that will provide information aboout automated vehicle operations and how human drivers interact with non-human ones. (Photo: Levi Hutmacher, The Michigan Engineer)
