Robotics

Could a robotic bike messenger ride the fast lane to autonomous delivery?

Could a robotic bike messenger...
The production version of the TWILL bot will most likely have a range of around 20 mi (32 km)
The production version of the TWILL bot will most likely have a range of around 20 mi (32 km)
View 5 Images
For the compartment size, the TWILL bot team is planning to match the trunk size of of a Toyota Prius
1/5
For the compartment size, the TWILL bot team is planning to match the trunk size of of a Toyota Prius
The team has built a prototype of the TWILL bot
2/5
The team has built a prototype of the TWILL bot
At this stage the TWILL bot is definitely more whiteboard-on wheels than robo-bike messenger
3/5
At this stage the TWILL bot is definitely more whiteboard-on wheels than robo-bike messenger
This kind of all-electric delivery robot would be mechanically simple
4/5
This kind of all-electric delivery robot would be mechanically simple
The production version of the TWILL bot will most likely have a range of around 20 mi (32 km)
5/5
The production version of the TWILL bot will most likely have a range of around 20 mi (32 km)
View gallery - 5 images

When picturing the autonomous delivery services of the near-future, you'll likely imagine drones buzzing overhead with packages in tow. But some ground-based robots are making a push into this area too, like the six-wheeled delivery droid recently dispatched to a customer's home with a food order onboard. Taking a similar approach is the team at LastMileRobotics, who is building a robotic courier based on a tried-and-true methods of the good ol' fashioned bike messenger.

Taking a two-wheeled approach to autonomous delivery could offer the same traffic-avoiding, cost-saving benefits of other concepts, such as those from the aforementioned Starship Technologies and airborne variants from Amazon and Flirtey. But the so-called TWILL bots (two-wheels in line locomotion) would be skinny and agile enough to roll up to your door. They could also boast a bigger range than drones, and lessen the risk of collisions with pedestrians and other vehicles.

"TWILL bots stand tall, on the scale of cars," LastMileRobotics co-founder Chris Tacklind explains to New Atlas. "This allows them to be seen as just another vehicle. But they are skinny, so they don't take up a lane of traffic. The big wheels allow them to traverse large cracks and even curbs. Their agility allows them to navigate pretty much anywhere people walk."

This kind of all-electric delivery robot would be mechanically simple
This kind of all-electric delivery robot would be mechanically simple

That's the idea anyway. The team has a lot of work to do before we actually see these things on the streets, but it says that in terms of basic mechanics they've got the big pieces in place. Kind of like a bicycle, the vehicle travels with two wheels in line, and when it comes to a stop, it turns one wheel to give it the stability needed to stay upright.

"Bicycles actually have a carefully selected geometry to make them self-stable at speed," Tacklind says. "We use an unstable neutral geometry that gives us higher control authority. We use tiny corrections in the steering angle to control the 'roll' of the vehicle. This actually harnesses the forward momentum to keep it upright. When stopped it is like a 'fixie bike.' With one the wheel turned, it balances like half of a Segway. When executed properly, you can barely see it move."

This kind of all-electric vehicle would be mechanically simple, Tacklind says, with no gyroscopes, kick stands, belts, chains or gears, making for an expected life of 250,000 mi (402,000 km), with the ability to handle terrain like a mountain bike (and possibly one day even tackle stairs).

The production version will most likely have a range of around 20 mi (32 km), Tacklind tells us, something that will be determined by the battery selection. For the compartment size, the team is planning to match the trunk of a Toyota Prius and give the robot a width of 17 in (43 cm), wide enough to handle the largest standard Amazon boxes. Top speed will likely be limited to 25 mph (40 km/h), a figure hoped to comply with common electric vehicle regulations.

The team has built a prototype of the TWILL bot
The team has built a prototype of the TWILL bot

The team has built a prototype, though it must be said, at this stage the TWILL bot is definitely more whiteboard-on wheels than robo-bike messenger. It stands one meter tall and 4.5 cm wide (39 x 1.7 in) and the team has used it to deliver manila folders indoors, but it is now making moves to get the robot out of the lab and into the world, though there are external factors that could impact its success.

"We have demonstrated basic functionality, so there is plenty of work do do," says Tacklind. "We are reaching out to academic groups and the open source community to push forward too. Ultimately, we will be limited by regulatory concerns which are being pushed by the self-driving car industry."

Part of that development work will also include the self-driving tech. "At this point we are 'autonomy agnostic'," Tacklind explains. "We don't want to be wed to our own autonomy solution. We'll use the best technology available for each market. There are plenty of vendors supplying autonomy hardware and software."

Source: LastMileRobotics

View gallery - 5 images
6 comments
Bob Flint
Have they tried this thing outside, seems that even a slight breeze, or incline, loss of traction, will play havoc with the continual power requirements of this half-baked idea.....
Doug Nutter
Automation and robotics, combined with artificial intelligence will eliminate many of the jobs presently being done by manual labor. The buggy whip manufactures are in full transition mode and the new jobs created will be more technological in nature. Our present education system doesn't begin to address the needs. My question is; what do we do with the students and people who are not capable or not willing to learn those things necessary for a living wage. Welfare is not the answer. It does not provide for one's self-respect, and we may face the situation of too few producers and too many takers. Do we go back to WPA days of the thirties? To let people live in a sub-standard manner invites violent revolution, so that's out of the question. Do we switch over to a guaranteed minimum income and tax the robots? Do we become a society of makers and artists? That has possibilities, or maybe a combination of both. I would like to see some serious discussions of the issues that we face. It's likely to come from Progressives who see the storm coming. Some Conservatives will, but many will refuse to address the changes that face us in very short order.
Nik
As a life long biker, I have serious doubts whether this device will manage to function on real roads, with all their defects, and dangers. What about in winter, with wind, rain, ice and snow? What about the effects of air turbulence caused by passing trucks, that can destabilise other full size vehicles. or other vehicles with impatient drivers, that may just ram this device out of the way, when there is no rider to consider? This seems to me to be a fun exercise, but with no real future, commercial or otherwise.
f8lee
@Doug Nutter - good points, all - and society will have to figure out what to do sooner than later. But as for this "bike" - for delivery purposes it seems to me the Boston Dynamics "Handle" wheeled & legged robot (seen elsewhere in this same issue of Gizmag, er, New Atlas) has this thing beat hands down.
xs400
So no one wants a cheaper, better and older way of doing this? Pneumatic tubes, anyone? But I guess that would be cheap, and no one would need all that tech.
Daniel Harbin
I can see these vehicles and other autonomous vehicles birthing a new game. A game where the object is to kill the robot, to knock it off course or over. I envision whole sites devoted to this new game and ways to beat any countermeasures or surveillance. Point totals and leader boards. Having ways to verify kills. Boy what fun.