Nov 8 2016
An autonomous mobility scooter and related software were designed by researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), the National University of Singapore, and the Singapore-MIT Alliance for Research and Technology (SMART). (Credit: Autonomous Vehicle Team of the SMART Future of Urban Mobility Project)
Over 100 visitors who attended MIT’s 2016 Open House rode on an autonomous mobility scooter. The ride is a trial of the software developed by researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT, the National University of Singapore, and the Singapore-MIT Alliance for Research and Technology (SMART).
The same software and sensor configuration have been used in the trials of autonomous golf carts and cars, and the latest trial marks the completion of the demonstration of a comprehensive autonomous mobility system. In principle, a mobility-impaired user can use the scooter to move across the hall and through the lobby of a building, use a golf cart to move across the parking lot and pick up an autonomous car on the road.
The latest trial proves that the control algorithms of the researchers work both indoors and outdoors.
We were testing them in tighter spaces. 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.
Scott Pendleton, a graduate student in mechanical engineering at the National University of Singapore (NUS) and a research fellow at SMART
There are many layers of software in the system developed by the researchers. These include low-level algorithms that allow a vehicle to immediately respond to changes in the surroundings, such as a pedestrian crossing the path; localization algorithms that are used by the vehicle to find out its location on a map; map building algorithms that help in the construction of the map; route-planning algorithms; scheduling algorithm that is used for assigning fleet resource, and an online booking system that enables people to book rides.
Uniformity
There are many benefits to using the same control algorithms for all kinds of vehicles such as cars, golf carts, and scooters. One such benefit is that, performing reliable analyses becomes more practical.
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. That’s useful for verifying that this multilayer complexity is correct.
Daniela Rus, the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT and one of the project’s leaders
Moreover, software uniformity enables the transfer of information acquired by one vehicle to another. For example, before it reached MIT, the scooter was tested in Singapore where it used the maps that were developed by the autonomous golf cart.
On the same note, Marcelo Ang, an associate professor of mechanical engineering at NUS and co-leader of the project, says that the researchers are currently incorporating machine-learning systems to their vehicles, so that the interactions with the surroundings can improve the performance of their control algorithms and navigation. “Once you have a better driver, you can easily transplant that to another vehicle,” says Ang. “That’s the same across different platforms.”
Software uniformity allows more flexibility to the scheduling algorithm in the allocation of system resources. In the absence of a golf cart a scooter can take a user across a park; a golf cart can fill in for a car, for a trip on back roads.
Talking about the system, Dan Ding, an associate professor of rehabilitation science and technology at the University of Pittsburgh says, “I can see its usefulness in large indoor shopping malls and amusement parks to take [mobility-impaired] people from one spot to another.”
Changing Perceptions
The scooter trial at MIT’s Open House also showed that the modular software and hardware system can be easily deployed in a new context.
It’s extraordinary to me, because it’s a project that the team conducted in about two months. [MIT’s Open House happened at the end of April] “and the scooter didn’t exist on February 1st.
Daniela Rus, the Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT and one of the project’s leaders
The design of the scooter and the outcomes of the trail have been presented in the first week of November, 2016 at the IEEE International Conference on Intelligent Transportation Systems. You Hong Eng, head of the SMART autonomous-vehicle project, and four other researchers from both SMART and NUS joined Rus, Pendleton, and Ang on the paper.
The outcome of a short user survey conducted by the researchers during the trial were also reported in the paper. Prior to riding the scooter, users were asked to rate the safety of autonomous vehicles on a scale of one to five; they were asked to answer the same question again after their rides. The average safety score was increased from 3.5 to 4.6 after the ride on the autonomous scooter.