Researchers have developed a lightweight mapping algorithm that reduces memory use and boosts the scalability of autonomous mobile robots.
Researchers at Northeastern University have developed a new algorithm that significantly improves the efficiency of mobile robot navigation .Designed to reduce the heavy memory demands of autonomous operation, the breakthrough enables robots to move and make decisions using fewer computational resources.
Named Deep Feature Assisted Lidar Inertial Odometry and Mapping , the algorithm paves the way for more practical and scalable deployment in real-world environments.Researchers claim the new 3D mapping approach is, in some cases, 57 percent more efficient than leading methods in the industry.Smarter robot navigationDelivery robots are increasingly becoming common, as they autonomously navigate city streets and neighborhoods.For efficient operation, these robots need a variety of sensors and sophisticated software algorithms. Lidar sensors are essential because they use light pulses to detect distances, allowing the robots to carry out simultaneous localization and mapping, or SLAM.The algorithm challenges the notion that more data equals better outcomes. Although efficient, this method uses a lot of memory to create and store precise maps, making it extremely resource-intensive. A considerable amount of computational strain can result from a memory load surpassing 10 to 20 gigabytes. This increasing demand constrains the robots’ capacity to operate over long distances or periods.Aiming for a solution to this, new research was conducted by a Northeastern doctoral student named Zihao Dong, under the guidance of Michael Everett, an electrical and computer engineering professor at the University. They developed a more efficient 3D mapping approach that could significantly reduce the computational burden on autonomous mobile robots. The new algorithm is up to 57 percent less resource-intensive than current leading methods, marking a major step in streamlining robotic navigation.Optimized robot visionDFLIOM builds on an earlier approach known as Direct Lidar Inertial Odometry and Mapping , which combines lidar sensors with inertial measurement units to map environments in three dimensions. While both techniques use similar technologies, DFLIOM introduces a novel scanning method that selectively processes only the most essential data points.This lowers the quantity of data needed and, in certain situations, increases the accuracy of the mapping. The study doubts the widely held business notion that performance improves with more data. The researchers contend that algorithms may become overwhelmed by too much data, resulting in slower processing and worse efficiency.Instead of depending on sensors gathering more data, the team developed more intelligent algorithms to recognize and use only the most pertinent data. With no memory or processing constraints, this change may allow mobile robots to work longer, quicker, and more efficiently in real-world settings.The new system is less resource-intensive than leading methods. The researchers utilized Northeastern’s Agile X Scout tiny robot, which came with an autonomous kit, an Ouster lidar sensor, a battery pack, and an Intel NUC tiny PC, to test their algorithm. The robot generated 3D maps of several outdoor locations on the University campus, including Centennial Common, Egan Crossing, and Shillman Hall, demonstrating the algorithm’s ability to perform in varied and complex environments.Researchers say the work represents a key advancement in developing practical, scalable autonomous systems. The details of the team’s research are available on GitHub.
Autonomous Autonomous Robots Deep Feature Assisted Lidar Inertial Odometry And Delivery Robots Direct Lidar Inertial Odometry And Mapping Navigation Northeastern University Robot Robot Navigation Sensors SLAM
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
The IRS must get smaller — and smarterPolitical News and Conservative Analysis About Congress, the President, and the Federal Government
Read more »
Run Smarter, Not Harder: Nike’s Running Shoe Finder Takes the Guesswork Out of Choosing SneakersA market-leader in the wellness space for over a decade, Well+Good is a digital publication and community dedicated to informing and empowering you to live a well life, inside and out.
Read more »
Lawrina and eForms: Comparing Features for Smarter Legal Document UseLawrina and eForms: Comparing Features for Smarter Legal Document Use
Read more »
Smarter Choice for Your Next VacationWhen it comes to planning a memorable spring or summer getaway, more travelers are thinking outside the box—or rather, outside the hotel room.
Read more »
Modeling method provides smarter way to predict customer demandResearchers have developed a new forecasting model that helps companies more accurately estimate how many customers are interested in a product -- even when key data is missing.
Read more »
The Smarter Web Company’s UK IPO To Include Retail Access And Bitcoin Treasury PlanUK retail investors can join The Smarter Web Company IPO as it integrates a bitcoin treasury strategy into its public market debut.
Read more »