Autonomous technology is making significant changes in several important areas of our lives, such as transportation, construction, infrastructure inspection, mining and utilities. According to Merriam-Webster, autonomy is “the state of existing or acting separately from others,” so autonomous technology refers to technology that functions independently, without receiving step-by-step instructions from humans.
To truly leverage the capabilities of the diverse range of sensors and devices currently available, autonomy is a necessity. People alone are unable to process the large volumes of data being produced, let alone interpret the results and make good decisions in a timely manner, whereas autonomous systems consisting of integrated hardware, software and custom algorithms greatly facilitate collection, processing and decision making for many applications.
Experimentation with autonomous technology has been ongoing for decades, with military applications being among the first developed to limit soldiers’ exposure to dangerous situations (e.g., drones and surveillance robots). More recently, civilian applications are gaining traction, with autonomous vehicles being one of the most high-profile areas of major investment and other industries, such as mining and utilities, already making use of autonomous technology in a variety of time- and cost-saving activities.
More Powerful Solutions Through Integration and Collaboration
The company AutonomouStuff LLC (AStuff) was founded in 2010 to focus on mobile robotics for transportation applications, although the range of uses being explored today goes far beyond that. By anticipating what basic functionality is needed to support the development of autonomous products, AStuff is able to offer packaged technology, such as LiDAR, RADAR, GPS, cameras, etc., that already works together. Instead of starting from scratch, a company with an idea for a new product leverages the research already done, and saves time and money by working with AStuff early on in its development process.
“The impetus for starting AutonomouStuff was my observation of a large gap in the supply chain that could significantly impact the future of transportation,” says Bobby Hambrick, founder of AutonomouStuff. “To expedite the creative process, we created a place where the world’s best researchers can come learn about the latest technology in the industry, and where they can speak with people who can help them with their projects.”
AStuff works with more than 30 vendors of equipment and software, including wireless control systems, LiDAR and RADAR sensors, and GPS/IMU/FOG. Integrating the capabilities of individual components results in a more powerful solution for the end user. AStuff also offers an automated research development platform that starts with a fully integrated drive-by-wire vehicle in which traditional mechanical control systems have been replaced with electronic control systems. A user then selects additional perception kits and middleware that form a custom foundation for their specific research and development needs. There is also the PolySync software platform that helps developers build, test and deploy applications quickly by selecting from a suite of APIs.
“Our goal is to find the best products available and make them work together, which allows our customers to focus on what they do best,” Hambrick says. “In some cases this saves them years in development time, which can equate to millions of dollars.”
One example is NVIDIA, he says. “Its CEO, Jen-Hsun Huang, recently gave a keynote speech at NVIDIA’s GPU Technology Convention, which showcased their advancements in deep learning for the future of automated driving. By using the AutonomouStuff platform, they were able to begin their innovative research and do amazing things in a very short period of time. That time advantage is nearly priceless for NVIDIA, especially considering the increasing competition in this field.”
Technology Transferable to Other Applications
All of the major auto manufacturers pursuing research in autonomous cars are customers, so automotive-related products are AStuff’s primary focus. However, other mapping and surveying applications also benefit from the integration of hardware and software provided by a variety of vendors. In addition to automated driving research, autonomous technology helps automate mapping and surveying in areas such as mining and agriculture, and military and naval operations.
“The PolySync software platform developed by Hambrick was originally intended for automotive applications, but it is really appropriate for a variety of customers with related needs,” says Terry Lamprecht, applications engineer at AutonomouStuff. “For every new application, we estimate that between 75 and 90 percent of the work is devoted to the software framework. PolySync saves programmers time by providing basic functions like time synchronization, parsing of data and extraction of data in a user-friendly format, which are required by most applications. We also provide drivers that work with most standard sensors and provide an electronic control unit (ECU), so in a very short period of time our customers can be writing and using apps. On day one, they can work with a variety of data without going through years of R&D.”
In recognition of the importance of software in the development of new autonomous applications, AStuff spun-off its software development group. Now Hambrick, based in Portland, Oregon, focuses on APIs in C or C++ that allow customers to pull in data from multiple sensors and work with that data seamlessly.
“Customers want to integrate a variety of different sensing modalities to provide redundancy and ensure detection in all environmental conditions,” Lamprecht says. “AutonomouStuff works as a value-added distributor for other vendors. We deal individually with each customer to find the best combination of technologies at the best price. We also offer first-level support with engineers trained in using all of the products. Our staff regularly goes on-site to train and help with operational issues, for example with RADAR and vision systems, and our software experts work on site to help with integration issues.”
One example of an AStuff vendor is Real Earth Inc., based in Pittsburgh, that uses advanced proprietary algorithms embedded in a processing board. The end result is a navigable 3D model, generated in real time, collected by any method: terrestrial, mobile, indoor or unmanned aerial system (UAS). The user doesn’t need existing maps or GPS, and the equipment is simple and low-cost.
“Real Earth uses advanced simultaneous localization and mapping (SLAM) algorithms together with LiDAR and IMU data to rapidly and interactively provide 3D maps at the point of work with no post-processing needed,” says Kevin Dowling, CEO, Real Earth. “There are many applications with great potential, including mining and construction, real estate development, tax assessment and zoning, as-built surveys and much more. Our systems are also useful for directing material-handling systems, providing information to first responders and calculating stock-pile volumes. Our system provides the world’s best mapping and localization system for many applications.”
AStuff is also a distributor for YellowScan, a company that uses an existing LiDAR scanner and integrates it with GPS/IMU, a battery and a processing platform on a UAS. Applications for UAS-carrying scanners include industry and infrastructure inspection such as wind turbines, bridges, telecommunication towers, pipelines, oil and gas refineries, mines, landfills and power lines. Depending on the level of accuracy required and the size of the area of interest, UAS are viable alternatives for obtaining topographic data combined with high-resolution aerial photography, video, thermal data, etc.
Autonomous technology is likely to reach into all aspects of our lives in the next 10to 20 years. “We are working in a time when we’re no longer restricted by the amount or type of data that we can process instantly,” Hambrick says. “As the technological hurdles are overcome, it becomes more about resources, time and money. First you have to have a product that people care about. Once you have a clear focus of what the end product looks like, it just takes a great plan and great people to get the job done.”