Provides a case study of a way of conveying intent on the road for autonomous vehicles.
You walk on the sidewalk and arrive at the crossroad at the same time as Uber. The driver raises his hand and leaves the steering wheel, waving his hand for you to pass. A minute later, you reach the second intersection, just after the other car, the driver raises his hand and tells you that they will go first.
But when you reach the next intersection, a self-driving taxi stops. How do you know what its plan is? Should you go first or does its artificial intelligence brain think the car should go first?
The entire self-driving world is asking a question: How does an autonomous car tell people around it what it is thinking?
situation
Currently, most companies have not solved this problem. Alphabet’s Waymo and GM’s Cruise have never mentioned this publicly, which is strange for two companies that put safety above profitability.
Most companies care about what they can do in the future. The idea of ​​the startup Zoox from the ground up was to build an autonomous vehicle (AV), not to modify an SUV like Waymo or Cruise.
Bloomberg Businessweek: Zoox prototypes VH1, VH2, and VH4
"The light-emitting diode lines on the front and back of the vehicle will send signals to other drivers... Its directional audio system will beep or dim to tell pedestrians on the crosswalk that vehicles have seen him." Bloomberg Businessweek's Ashlee Vance wrote.
By using external lighting, Zoox attempts to establish a "lighting language". The question is, "How do you make the public aware of this new standard?" With the introduction of new colors, blinking patterns, and even lights on cars, people are likely to panic and don't know what's going on. The same goes for the "directional sound system" mentioned by Vance. Will people understand the difference between aggressive and peaceful tone? It is a fact that people resist change. For example, when the United States tried to change from the imperial system to the metric system, almost no one wanted to learn a new measurement system, so this work failed. The road signs in a few towns have changed from a few miles to a few kilometers, but they never surpassed this sign.
drive.ai: A passenger enters one of the taxis in Frisco, Texas
Drive.ai is one of the few companies that have public taxi services. In fact, they are the only company that allows their cars to communicate with pedestrians.
Their large orange truck provides a taxi-hailing service in Texas, with boxes on each side of the truck. drive .ai only displays one language, and the text can only be read from certain angles, so it's a design problem.
Take my hometown as an example: 800 languages ​​are spoken in New York alone; English is not the main language of many New Yorkers. Using text to help convey intent will lock drive.ai to a specific demographic.
In the United States, approximately 21 million people have visual impairments. In order to create maximum visibility, large-scale comparisons must be made. As we all know, in broad daylight, the screen does not provide such a function, which makes glare a problem.
Nissan Research Center: Slideshow shown in Google design [Autonomous]
Nissan hasn't announced their cars yet, but a lot of research has been done. Combining the plans of Zoox and drive.ai, Nissan hopes to use external lighting and screens. Coming back to the question of contrast, these lights and screens must be very bright, especially in the daytime, in order to improve the vehicle's own intentions and pedestrian awareness. This form of communication between cars and pedestrians—language—needs to be stronger.
When I saw this, I had to ask myself, "What kind of solution does not require learning, but can be effective for everyone in the community?"
spark
New York City is notorious for its energy. I was born and raised in New York, and everyone walks fast. Our favorite sentence is, "If it is red, it is green."
Citi Bike is very suitable for New York, allowing New Yorkers to get where they need to go faster than before. Although the main mode of transportation in this city is the subway, Citi Bike is becoming more and more popular in New York.
Thanks to Citi Bike and Blaze
Last year, Motivate installed a laser device in the bicycle basket of Citi Bikes, projecting the bicycle icon onto the sidewalk 15 feet ahead of its own (those pedestrians who would not look up at their phones before, but Will look up at you). Not only is this visible at all times of the day, it is the perfect early warning system for New York City’s nascent bicycle lane system.
Although this laser has its drawbacks-it is very unstable and a bit small on the sidewalk-but it can trigger my thinking. "Can this be adapted to AV?", "Will this change the color?", "What will happen if these lasers are connected to motors and gimbals?" All thoughts immediately came to my mind.
Proof of concept
When I was playing on Sketch, I realized that cars can use this technology like a bicycle, and in a better way. Enter Wave: The car can project a parking icon on the sidewalk to clearly convey its intention, instead of turning the light red to signal pedestrians that it should stop.
Connected to the motor, when the car is driving towards it, the protrusion can even stay in place, eventually "passing" the protrusion itself.
To start an accurate prototype, the first stop I did was the noun project, the one-stop shop icon.
Noun item stop sign result
Since most of the icons are octagonal, however, I started to design the way cars projected stop signs on the crosswalk. I think this way, a pedestrian who wants to cross the road will see that the car will drive first, and they should stop before crossing the road.
"I do not understand"
When I introduced Wave to my friends, I was surprised to find that their faces were confused. "Wait, is this car saying to stop, or is it telling pedestrians to stop?"
I realized that there is a fatal flaw in my design: it is not neutral to the different languages ​​of drivers and pedestrians. When the car sees the red octagon, it will give in, and the pedestrian will only stop when it sees the red hand. What is more confusing is that when the car sees a green circle, the car is gone, and when they see a white crutches, pedestrians will pass.
Obviously, a clear distinction needs to be made between the projection that signals the car's intention and what is required of pedestrians.
And, in order to unify the differences between the meanings of colors and icons around the world, I couldn't stop working on the noun project. I need to search the Internet for the commonalities between the transportation icons of various countries to create colorful icons that everyone can understand.
Next step
After research, I found that the AV needs to communicate with pedestrians, and the car needs to project icons familiar to pedestrians when moving. In the crosswalk scenario, the car projects a green walking icon to tell pedestrians that they can cross. After crossing, the car can project a red hand to signal to others that it is now his turn.
Another advantage of Wave is that, unlike the current second two drivers, who wave or raise their hands when driving today, these icons can be projected five times as long. This not only increases visibility, but also maximizes the safety of self-driving cars.
Turn
The United States is home to one of the strangest traffic practices ever: four-way parking. If you think about it, four-way parking is a symphony of different moving parts. The politics between cars and pedestrians, cars and cars are very important for AVs that observe and take action. Melissa Cefkin explained this very well.
https://youtu.be/5hLEiBGPrNI
Remember the reference to the Zoox project earlier in the article? Cefkin actually owns the patent.
The biggest pain point for cars and pedestrians is that they don’t know what their opponent’s next move is. Pedestrians will think "Do I have enough time to run across the street? Drivers will think "Do I get to the intersection first or do they get to the intersection first?" "
Wave can also solve this problem. With lidar, that is, the eyes of any self-driving car (or the top slender object), AV can know in advance the route outside of 2-3 football fields. Cars can easily take advantage of the four sites.
Wave enters an intersection with the purpose of turning left and projecting their planned path a few feet ahead before they execute the turn.
This allows pedestrians and other drivers to see the intentions of self-driving cars and where they plan to go, so that they have more knowledge to take action. They will not stop the pedestrian in front of the car, but can stop one foot from the projected path of the car, knowing that its turning radius is because of Wave.
Wave fundamentally reduces unexpected conflicts, simplifies intersections, and makes self-driving vehicles safer. Even before this problem occurs, it can change the way pedestrians and cars interact.
Waymo's Firefly and Wave technology superimposed
What should we do now?
Wave can only go so far on my computer. The next step in the process is obviously to test it in the real world. Sadly, this is where my expertise has stagnated. Not only do I not have a self-driving car to play, I don’t even have a virtual car to experiment with.
The limitations of my computer are not just limited to testing, they also prevent me from getting feedback from people outside my circle. This is why I am in the media. I hope that researchers can help me take Wave to the next level (wink, wink).
The race for self-driving cars is at the highest point in history. Companies like Waymo have begun to want to use real people to test programs without a driver. Companies such as Uber have retreated due to industry pressures. However, there are companies like Zoox that can fall from the sky to replace the fallen.
If you have the resources to help test Wave, I would be happy to talk about it. If you have cash to sponsor this patent, please email me immediately.
At present, I will insist on conducting more case studies in the field of autonomous driving.
Zhejiang Kaimin Electric Co., Ltd. , https://www.ckmineinverter.com