Self-Driving Scooter Helps Mobility-impaired People Get Around
The self-driving scooter uses the same sensors and software that had been used in previous autonomous car and golf cart tests.
Researchers at MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and the National University of Singapore have developed a self-driving scooter that uses the same sensors and software that had been used in previous autonomous car and golf cart tests.
Here’s the idea. Self-driving cars can only transport a mobility-impaired person part of the way - from their home to a mall, let’s say. How would that person get around the mall? That’s where a self-driving scooter comes into play. The researchers say the self-driving scooter works just as well indoors as it does outdoors.
“We were testing them in tighter spaces,” says Scott Pendleton, a graduate student in mechanical engineering at the National University of Singapore (NUS) and a research fellow at SMART. “One of the spaces that we tested in was the Infinite Corridor of MIT, which is a very difficult localization problem, being a long corridor without very many distinctive features. You can lose your place along the corridor. But our algorithms proved to work very well in this new environment.”
The self-driving scooter has several layers of software: low-level control algorithms that enable a vehicle to respond immediately to changes in its environment, such as a pedestrian darting across its path; route-planning algorithms; localization algorithms that the vehicle uses to determine its location on a map; map-building algorithms that it uses to construct the map in the first place; a scheduling algorithm that allocates fleet resources; and an online booking system that allows users to schedule rides.
Daniel Rus, one of the project leads, says using the same control algorithms for all self-driving vehicles, whether it be scooters, golf carts, or cars, has several advantages. One is that it becomes much more practical to perform reliable analyses of the system’s overall performance.
“If you have a uniform system where all the algorithms are the same, the complexity is much lower than if you have a heterogeneous system where each vehicle does something different,” says Rus, the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT. “That’s useful for verifying that this multi-layer complexity is correct.”
Before riding the scooter, users were asked how safe they considered autonomous vehicles to be, on a scale from one to five; after their rides, they were asked the same question again. Experience with the scooter brought the average safety score up, from 3.5 to 4.6.
The scooter trial also demonstrated the ease with which the researchers could deploy their modular hardware and software system in a new context.
“It’s extraordinary to me, because it’s a project that the team conducted in about two months,” Rus says. MIT’s Open House was at the end of April, and “the scooter didn’t exist on February 1st,” Rus says.
In a similar project, the University of Washington Bothell is building a self-driving bicycle that can travel 30 MPH with a 15-mile range. The team claims the self-driving bicycle “could get riders from A to B faster than a self-driving car, and at a tenth of the cost.”
he self-driving bicycle is based on a stack of five Arduinos with no operating system. The open-source documentation and C/C++ code can be found on GitHub. The self-driving bicycle has object recognition that’s based on an array of sonar range finders. Here’s more from Folsom on how the self-driving bicycle sees its environment:
“Localization combines GPS with dead reckoning. It uses a fuzzy filter to reconcile the two estimated positions. Dead reckoning is based on speedometer and magnetometer. In the future, localization may be extended to include accelerometer, optical odometry and visual lane edge detection. A Raspberry Pi handles the visual tasks.