Southwest Research Institute, a developer of software and systems integration solutions for automated vehicles and robotics, has deployed an automated shuttle at its 607ha campus in San Antonio, Texas.
Capable of ferrying up to 14 passengers, the shuttle features algorithms, sensors, cameras and software that SwRI developed through internal research.
“It is rewarding for our engineers to take the very best technology that SwRI has developed to serve our clients and then embed it into a showcase vehicle that has a practical purpose in our backyard,” said Ryan Lamm, director of SwRI’s Applied Sensing Department.
Using SwRI’s Ranger system, the vehicle has been programmed to autonomously drive unique routes around the grounds. Ranger is a localization tool that uses a ground-facing camera and automation software to precisely maintain its position to within 2cm on a given route. During operation, a human driver sits behind a steering wheel as an additional safety measure.
“Driverless, fully automated vehicles are still several years away, but this shuttle proves that we are well on our way to such a future,” added Steve Dellenback, vice president of SwRI’s Intelligent Systems Division.
The shuttle is classified as a low-speed vehicle (LSV), operating at less than 50mph, making it ideal for closed campuses, such as the SwRI headquarters, and roads with lower speed limits. It uses a campus map created in the Ranger system with features such as intersections, lanes, stop signs and crosswalks. Routes are dynamically selected by the operator along the map.
It also features artificial intelligence (AI) to classify roadway signs, pedestrians, vehicles and other objects. SwRI has programmed the shuttle with various driving scenarios such as sharing the road with other vehicles, detecting dynamic objects such as pedestrians and cyclists, and determining right of way at intersections.
“The shuttle collects data every time it is on the road, allowing us to continually refine its algorithms to improve reactions in various situations,” noted Alexander Youngs, an SwRI senior research engineer who led development of the shuttle.
The shuttle’s core functionality lies in SwRI’s autonomy stack, a suite of tools using proprietary machine learning algorithms, software and processing tools as well as cameras and sensors. Additional capabilities include the ability to integrate and share data with intelligent transportation systems and other connected and automated vehicles. The shuttle can be deployed in a convoy of similar vehicles, taking traffic and congestion into consideration for improved mobility.