One of the pioneering companies that helped spark the autonomous mobile robot (AMR) revolution is now turning its artificial intelligence capabilities toward automating another major workflow challenge - pallet handling. At this week's LogiMAT trade show in Stuttgart, Mobile Industrial Robots (MiR) unveiled its new MiR1200 Pallet Jack, which the company hails as the industry's most advanced AI-enabled autonomous pallet mover.

The MiR1200 represents the Danish AMR maker's biggest technological leap yet by incorporating advanced 3D vision, machine learning and a host of sensors to enable fast, reliable, and precise pallet pickup and delivery across warehouses and facilities.

"We believe the AI detection system in the MiR1200 is a significant improvement over older pallet handling technologies," said Mads Paulin, MiR's VP of R&D. "Our approach reduces pick-and-place cycle times, delivers best-in-class accuracy, and allows us to continuously update the system with new AI capabilities."

Solving the 'Pallet Problem' with AI While AMRs have made tremendous strides automating basic point-to-point material transport, pallet handling has remained a persistent challenge due to the irregular shapes, conditions and orientations pallets often appear in real-world environments.

Reliably detecting and manipulating undamaged, neatly stacked pallets is hard enough. But pallets strewn with debris, covered in shrink wrap or marred by cracks and damage require more advanced perception and decision-making capabilities.

That's where the computer vision and AI smarts of the MiR1200 Pallet Jack come into play. The robot's sensor stack combines RGB depth cameras and 3D lidar data, processed through NVIDIA's high-powered Jetson AGX Orin module.

Using simulation tools from NVIDIA Isaac, MiR trained its AI models on a massive 1.2 million synthetic and real-world images of pallets in various states and environments. This allowed the MiR1200's deep learning systems to grow adept at accurately identifying intact pallets ready for pickup while avoiding false positive detections of debris or irregularities.

"With simulation, we can validate the system's performance and reliability by testing variations like changes to floors, lighting conditions or pallet appearances," said Kevin Dumas, VP of Product at MiR. "It future-proofs the MiR1200 so we can continuously expand its AI capabilities."

The MiR1200 then dynamically modifies its route plan to navigate around obstacles and through tight spaces thanks to advanced sensor fusion and the NVIDIA Jetson's ample compute power. Safety remains paramount, with the robot slowing or stopping when required based on 3D environment mapping.

NVIDIA's "AI Plumbing" Fuels New Robotics MiR's strategic collaboration with NVIDIA illustrates how the chipmaker's push into robotics extends well beyond just high-performance GPUs and AI models.

As the new owner of MiR following its $275M acquisition of parent company Teradyne's robotics business units last November, NVIDIA is now integrating its full AI stack - processors, simulation tools, fleet management, and more - into advanced AMR products.

"NVIDIA recognized the value of MiR's complementary ecosystem of partners, OEMs, and end users when it comes to commercializing real-world AI robotics at scale," said Ujjwal Kumar, the former head of Teradyne Robotics now overseeing NVIDIA's AMR business. "We're combining their market leadership with our AI 'plumbing' to create smarter robots for the warehouse."

Simulation, compute horsepower and software integration are proving just as vital as the AI algorithms themselves in bringing advanced autonomy like the MiR1200 to life.

Beyond the technological capabilities, NVIDIA's deep pockets and robotics investment roadmap could propel its recently-acquired AMR business into the vanguard of AI-based warehouse and logistics automation over the coming years.

Demand Surging for Automated Pallet Movers If successful, the MiR1200 addresses a clear and growing need throughout supply chains and industrial facilities.

Pallet manipulation is consistently cited by existing MiR AMR customers as one of their biggest bottlenecks and opportunities for automation due to widespread labor shortages affecting warehouses and distribution centers.

"Customer demand for automated pallet handling seems insatiable right now," said Dumas. "The MiR1200 was built from the ground up to be a rugged, industrial-grade AMR pallet jack that works seamlessly in real-world warehouse environments where you find pallet damage, debris and other challenging conditions."

The MiR1200 is designed to integrate into existing MiR AMR fleets and workflows, according to Jean-Pierre Hathout, President of MiR. It can be managed through the company's MiR Fleet software for coordinating robot teams across multiple facilities.

Addressing labor gaps, improving throughput, and enhancing worker safety were all key drivers behind the platform's development.

"Safety is hardwired into the MiR1200's DNA," said Hathout. "Its sensor fusion of cameras and lidar provides precise detection of obstacles, overhead and all around. When combined with our proven safety practices, it offers a much safer alternative to manual pallet jacks and forklifts."

As one of the first major AI-based robotic offerings from MiR under NVIDIA's ownership, the MiR1200 Pallet Jack could just be the start of an accelerated roadmap blending AMRs with more advanced autonomy, perception, and general intelligence capabilities.

There's no shortage of materials handling challenges remaining to be automated across supply chains. With AI like the MiR1200 tackling long-standing roadblocks in pallet handling, warehouses are poised to become substantially more intelligent and automated in the coming years.