FedEx has introduced a groundbreaking two-armed robot named DexR, engineered to automate one of the most challenging tasks faced by the company’s human workers—loading delivery trucks with packages. This innovative robot leverages artificial intelligence (AI) to efficiently stack rows of differently sized boxes within a delivery vehicle, aiming to maximize cargo space.

The complexity of this task is not to be underestimated. “Packages come in different sizes, shapes, weights, and packaging materials, and they come randomized,” explains Rebecca Yeung, Vice President of Operations and Advanced Technology at FedEx. DexR relies on a combination of cameras and lidar sensors to perceive and assess the packages, followed by intricate planning to configure the available boxes into a compact, secure arrangement, all while avoiding damage to the contents. Yeung highlights the significant advancements in AI that now make such complex decision-making feasible. DexR is currently undergoing testing, with plans for wider deployment across FedEx’s operations in the future.

Despite the impression created by generative AI tools like ChatGPT, which have demonstrated AI’s versatility across various industries, handling objects in the real-world environment remains a formidable challenge for algorithms. Most industrial robots excel at performing highly repetitive tasks with unwavering precision but struggle with tasks that involve variability.

Progress is being made by roboticists. An increasing number of machines employ AI to recognize objects and determine optimal grasping methods. These algorithms are often trained within simulations, where errors bear minimal consequences before being transferred to real-world robots. However, transitioning from simulation to reality poses considerable hurdles.

Recent strides in AI algorithms, novel approaches to applying machine learning to robotics, and advancements in hardware and sensors are expanding the range of commercial applications for advanced robots. Matthew Johnson-Roberson, Director of the Robotics Institute at Carnegie Mellon University, notes that significant investments in domains such as autonomous vehicles, combined with continuous AI progress, are positioning robots for wider adoption in various workplaces. He envisions a forthcoming wave of commercial robotics.

FedEx’s DexR robot was developed by Dexterity, a startup headquartered in Redwood City, California, specializing in AI-driven robotic systems for diverse warehouse tasks. Dexterity CEO Samir Menon reveals that DexR employs generative AI to determine optimal stacking strategies for various box types while also utilizing AI for object identification and handling. However, the successful integration of these systems relies on meticulous engineering.

With the growth of e-commerce, especially driven by companies like Amazon, the realm of package handling has emerged as a dynamic frontier for robotic innovation. Amazon, for instance, is deploying thousands of advanced robots in its facilities to continually enhance operational efficiency in product storage and processing.

By Impact Lab