Revolutionary VORA Tech: Changing the Game for Off-Road Navigation SwRI Develops Cutting-Edge System for Military Autonomous Driving

Southwest Research Institute has created new technology for autonomous driving, specifically designed for military use with an emphasis on stealth and flexibility for clients in the space and agricultural industries.

The innovative system utilizes a pair of stereo cameras and unique algorithms, eliminating the need for lidar and active sensors.

The team at SwRI conducted in-house research and developed a suite of tools called Vision for Off-road Autonomy (VORA). This passive system has the ability to detect objects, map the environment, and simultaneously locate and navigate in off-road settings.

The VORA experts envisioned a camera-based system as an alternative to lidar, which uses active lasers to measure depth and distance. While lidar is highly reliable, its use may reveal the presence of hostile forces. Radar, another commonly used sensor, is also detectable. In contrast, GPS signals can be disrupted, and their accuracy can be limited in certain environments, such as canyons and mountains, which can impede agricultural automation.

"We recognized the need for more advanced passive sensing capabilities for our military clients but also discovered the potential for this technology in other industries such as agriculture and space research," explained Meera Towler, an assistant program manager at SwRI who led the project.

The VORA technology was initially developed for exploring planetary surfaces. In space applications, robots face challenges such as limited power, payload capacity, and intermittent connectivity. In these environments, cameras are a more practical option than power-hungry lidar systems.

To address various obstacles, the team created new software that utilizes stereo camera data to perform tasks that are typically accomplished with lidar. These tasks include localization, perception, mapping, and world modeling.

As a result of this research, SwRI introduced the deep learning stereo matcher (DLSM) tool, which uses a recurrent neural network to generate dense and precise disparity maps from stereo images. A disparity map highlights the differences in motion between two stereo images.

To assist with simultaneous localization and mapping, SwRI developed a factor graph algorithm that intelligently combines sparse data from stereo image features, landmarks, inertial measurement unit (IMU) readings, and wheel encoders to generate precise localization data. Factor graphs, also known as probabilistic graphical models, are used in autonomous systems to make comparisons between variables and make inferences.

"Our research in autonomy has applications in various industries, including military vehicles, commercial vehicles, and agriculture. We are thrilled to offer our clients a plug-and-play solution with a stereo camera integrated into a leading autonomy system," said Towler.

SwRI intends to incorporate VORA technology into other autonomy systems and test it on an off-road course at its San Antonio campus. The organization has prioritized safety and security in the development of autonomous vehicles and automated driving systems as the technology advances and becomes more suitable for commercial and government use.

David Lamy
David Lamy Author
David Lamy owns the Bachelor in Atmospheric Science Degree. He is associated with Industry News USA from last 2 years. With proficiency in his work, David obtained a strong position at Industry News USA and heads the Science section. “Weather forecasting” is the field of his interest. He bags total 5 years of experience in this field. Apart from his routine work, David loves to explore his cooking skills. He has participated in various cookery shows.