There has been a lot of talk recently about autonomous, self-driving cars being tested by Google and other auto manufacturers, but a group of researchers from the Massachusetts Institute of Technology (MIT) has been working on a lesser-known project to create a self-driving golf cart.
The team at MIT has partnered with a counterpart in Singapore and a technology company called Velodyne LiDAR Inc. to produce a fully autonomous golf cart, and the technology may soon be adapted to other types of vehicles, including neighborhood electric vehicles (NEVs) and full-size cars.
First Tests Began Last September
Although it has never been revealed who came up with the original idea for a self-driving golf cart, the team responsible is a partnership between researchers at MIT and the National Research Foundation of Singapore, whose members come from several different technology research universities in the country.
The Singapore-MIT Alliance for Research and Technology (SMART) kept their plans to develop an autonomous golf cart relatively quiet in the time leading up to the first public test runs, which occurred last September in a public park in Singapore. However, this past April as the Masters Tournament began in Augusta, Georgia, Velodyne LiDAR Inc. took the opportunity to announce it had partnered with SMART by providing key technology in the form of its real-time 3-D sensor system: the VLP-16 LiDAR Puck.
The test run of the custom golf carts included 500 unaffiliated participants who had used an online booking system to schedule pickups and drop-offs at 10 stops established at various locations around the park. The golf carts successfully ferried all 500 subjects through winding paths populated with a range of obstacles, including pedestrians, bicyclists, and wildlife.
Modified Yamaha Electric Golf Carts
Soon after plans were solidified, SMART went to work using two Yamaha electric golf carts as base models. Each was equipped with a state-of-the-art webcam and the VLP-16 LiDAR Puck from Velodyne, which is one of the most affordable and powerful sensors of its kind. The team had tried to use an earlier version that only included 2-D sensing technology, but initial tests were less than satisfying.
When the 16-channel, realtime 3-D VLP-16 Puck was released, Velodyne presented it to SMART, and soon afterward, the self-driving golf carts were ready for real-world testing. Using the new technology, the two custom golf carts were able to successful transport all 500 passengers through winding trails as the sensors fed information to the system’s computer, which processed that information to make choices concerning speed, acceleration, deceleration, positioning and stopping.
Dynamic Virtual Bumper
During the test run, the two custom golf carts were able to reach speeds of up to 15 mph as they jockeyed for safe positions through the garden paths. This was made possible by what SMART is calling a dynamic virtual bumper. Rather than being steered completely by a global positioning system (GPS) as most people assume, the sensor and camera on each cart creates an invisible sensing area in the shape of a cylinder that encompasses the vehicle’s planned trajectory. The exact width and length of this virtual bumper is determined by the vehicle’s speed.
When an object enters the dynamic virtual bumper, the sensors collect information and send it to the onboard computer so that it can instruct the drive system to slow down or maneuver until the object is no longer in the cylindrical bumper zone.
The online booking system used to schedule pickups and drop-offs is experimental in nature and is also being tested alongside the autonomous golf carts. The system successfully deployed the vehicles when and where they were needed, which included routing and rerouting on the fly to meet newly scheduled itineraries.
Test Riders Highly Satisfied
According to the surveys given to each passenger in the tests, 98 percent said they would definitely use the autonomous golf carts again, and 95 percent stated that they would return to the same park if the self-driving cars became a regular attraction.
SMART went on record stating that the team had taken a minimalist approach to the development of the self-driving golf carts. “The vehicles are instrumented, but they are not as heavily instrumented as the Google car,” said Daniela Rus, a professor in MIT’s Department of Electrical Engineering and Computer Science. “We believe that if you have a simple suite of strategically placed sensors and augment that with reliable algorithms, you will get robust results that require less computation and have less of a chance to get confused.”
Velodyne also made a statement regarding the autonomous golf carts. “Velodyne LiDAR has become the de facto standard for autonomous vehicles,” said Wei Weng, sales director of Velodyne Asia. “Whether on an actual golf course, a park with pathways or the open road, LiDAR technology is consistently reliable whatever the lighting or weather conditions.”