Bicycles

Autonomous tricycles could form the basis of urban taxi systems

Autonomous tricycles could for...
UWB engineers have developed a self-driving trike, that they hope will eventually lead to more affordable autonomous vehicles
UWB engineers have developed a self-driving trike, that they hope will eventually lead to more affordable autonomous vehicles
View 5 Images
Another model of self-driving trike the team experimented with, this one still containing a seat
1/5
Another model of self-driving trike the team experimented with, this one still containing a seat
UWB engineers have developed a self-driving trike, that they hope will eventually lead to more affordable autonomous vehicles
2/5
UWB engineers have developed a self-driving trike, that they hope will eventually lead to more affordable autonomous vehicles
The project has received a US$75,000 grant from Amazon Catalyst
3/5
The project has received a US$75,000 grant from Amazon Catalyst
The team put the self-driving trike through its paces by having it drive a circle shape 
4/5
The team put the self-driving trike through its paces by having it drive a circle shape 
Some of the team behind the project (Front row, from left):  Jeremy Bobotek, Dylan Katz, Aaron ConradBack row: Shyawn Karim, Tyler Folsom, Chase Skelton, Pengfei Zhu, Varsha Srivastava
5/5
Some of the team behind the project (Front row, from left):  Jeremy Bobotek, Dylan Katz, Aaron ConradBack row: Shyawn Karim, Tyler Folsom, Chase Skelton, Pengfei Zhu, Varsha Srivastava

Self-driving cars, trucks and buses might get the bulk of the headlines, but a team at the University of Washington Bothell (UWB) is developing a smaller kind of autonomous vehicle. With the aim of providing a relatively inexpensive alternative to owning an autonomous car, the team is creating a self-driving trike that may even open up the possibility of an automated ride-sharing network, like a bike version of Uber's or NuTonomy's proposed services.

The team, headed up by Tyler Folsom, has been experimenting with fitting autonomous systems into tricycle frames and this work culminated in August with a test that saw a bright orange recumbent trike drive itself in a circle. That modest command, entered via remote control, demonstrated the vehicle's ability to stop, start and turn itself to reach a destination, but Folsom says it's just a "baby step" on the way to deeper autonomy.

"I'm trying to shift the talk about self-driving cars to self-driving bicycles and making sure bicycles are part of the automation equation," says Folsom.

The outcome of that equation, the team hopes, is to eventually produce autonomous vehicles that are much lighter and more environmentally friendly than self-driving cars. With a targeted price tag of around US$10,000, ideally they'd be cheap enough to replace the family car or current public transport options. To keep that price down, the team is trying to maximize the efficiency of the electronics driving the trikes.

"We're using things much less powerful than a smartphone," says Folsom. "Part of the concept is that you don't have to spend as much money as the big car companies are spending. My contention is you don't need all that much processing power to make autonomy happen."

The team put the self-driving trike through its paces by having it drive a circle shape 
The team put the self-driving trike through its paces by having it drive a circle shape 

Reducing the required computational power may be easier to achieve if human error is removed from the picture by setting up a better autonomous infrastructure, which is a goal Folsom has been vocal about for years with his Elcano Project. Along with dedicated lanes for autonomous vehicles, he puts forward the idea of renewable energy-powered self-driving taxi systems, possibly with a fleet of velomobiles like Organic Transit's ELF, which could ferry people around cities without impacting too heavily on the environment.

"The big thing for me is the effect this could have on global warming," says Folsom. "If we can push transportation in this direction – very light vehicles – it's a major win for the environment. I want to have the technology that lets people make that choice if we decide, yes, by the way, survival would be a nice thing."

The project, which involves over 20 people, has received a $75,000 grant form Amazon Catalyst.

The team describes their work in the video below.

Source: University of Washington Bothell

UW Bothell Autonomous Tricycle

3 comments
Chizzy
Arcimoto has been saying all along that their goal is to build the vehicle first, then add autononomy. the advantage of their trike, is no pedaling. the advantage of a bicycle is its thrust to weight. the moment you start adding anything, that advantage begins to slip away. if i have to be aware enough to pedal, surely i can be aware enough to drive the thing, so i don't see the point of a human power autonomous vehicle.
byrneheart
I could definitely pedal and read at the same time. I do it on my exercise bike, why not in an autonomous cycle?
StWils
The only significant comparison between autonomous bikes & trikes and larger vehicles is the degree of crushing & compaction that will occur when they get hit. In no uncertain terms I think that people should control the vehicle they are riding in. While autonomous control sounds OK on a completely closed road space such as a rail line or an interstate highway it simply seems crazy to permit this on open local roads. Also, while the rail industry has had great success with Positive Train Control where this has been implemented that still cannot tell when a car, truck on pedestrian is on the tracks. You need an alert driver for that. Autonomous vehicles still need a somewhat closed controlled road. However, all that said, I can see value in an autonomous vehicle moving itself around in a "Go Fetch" mode so long as the vehicle controls are able to totally yield to all other traffic. This will still be somewhat annoying as slower vehicles are bypassed but still tolerable. A sufficiently brightly lit autonomous taxi service is a great idea. One of the basic structural costs for mass transit is the bus driver, relatively speaking the high cost of the bus is surpassed by high personnel costs. This establishes both a minimum size for a bus AND a relatively rigid schedule & route. An autonomous small, brightly lit vehicle could serve an area with considerable flexibility. Whether such drive controls will be capable of responding safely to stupid people, pets, deer, irrational drivers, the occasional falling tree, etc, is still a very uncertain situation.