NVIDIA Builds Isaac AMR Platform to Support $ 10 Trillion in Logistics Industry


DeepMap, Metropolis and ReOpt are improving the performance of fleets of autonomous mobile robots, given expectations that robot locations will increase nearly six times by 2025.

Manufacturing and fulfillment centers are extremely complex. Whenever new earphones or socks land on your doorstep within hours or a vehicle rolls off the assembly line, a labyrinth of magic happens with AI-controlled logistics.

Massive systems like this one are constantly in motion. Robots travel miles of corridors to roll millions of products together to support teams of people in motion. Obstacles are omnipresent.

We’re introducing the Isaac Autonomous Mobile Robot (AMR) platform to optimize operational efficiencies and accelerate the deployment of AMRs. Isaac AMR extends NVIDIA Isaac’s ability to build and deploy robotic applications, bringing mapping, location analysis, and fleet optimization to NVIDIA EGX servers.

Industrial facilities of this type can be as large as city blocks or stadiums. They are constantly being reconfigured or scaled to meet current product requirements. Path planning and rerouting for autonomous robots must be done in lockstep.

On an industrial scale, even small routing optimizations in the logistics industry can save $ 10 trillion worth of billions of dollars.

According to Interact Analysis, the deployment of autonomous mobile robots is estimated to reach 53,000 locations by 2025, up from 9,000 in 2020. Meanwhile, supply chains are struggling to keep up with increasing e-commerce amid labor shortages and COVID-19 restrictions.

One hurdle is the ability to develop high-precision robot maps quickly and autonomously. They need to be continuously updated as operations scale or fluctuate. And increasing situational awareness of the changing environments for mobile robots and continually re-optimizing routes – while developing new skills through simulation – is paramount to operational efficiency.

Isaac AMR is the result of years of research and product development at NVIDIA. The framework is available on the NVIDIA NGC software hub and within the NVIDIA Omniverse platform and will initially fall back on Metropolis and ReOpt and soon on DeepMap and other NVIDIA technologies.

Scaling operations with Isaac AMR

The AI ​​and computing challenges of autonomous mobile robots for manufacturing and fulfillment are not too different from those of autonomous vehicles.

Obstacles and people must be avoided. Goals have to be achieved. Thousands of sensors powered by GPU-accelerated algorithms help fleets of autonomous robots solve the traveling salesman problem – finding the shortest route between multiple destinations – in real-time, with ever-changing industrial workflows.

The Isaac AMR platform uses NVIDIA Omniverse to create digital twins of the facility where AMRs are deployed. NVIDIA Isaac Sim (built on Omniverse) simulates the behavior of robot fleets, humans and other machines in the digital twins with high-fidelity physics and perception. It also enables synthetic data generation for training AI models.

Isaac AMR consists of GPU-accelerated AI technologies and SDKs including DeepMap, ReOpt and Metropolis. These technologies are securely orchestrated with NVIDIA Fleet Command and made available via the cloud.

DeepMap offers mapping advances

NVIDIA’s recent acquisition of DeepMap also brings advances in autonomous vehicle mapping for the AMR industry.

AMR deployments can access the cloud-based SDK of the DeepMap platform to accelerate robotic mapping of large facilities from weeks to days while achieving centimeter accuracy.

With the DeepMap Update Client, robot maps can be updated in real time as often as necessary. And the DeepMap SDK delivers layers of intelligence to the maps by adding semantic understanding – so robots can identify the objects that represent pixels and know whether or not they can move in one direction. It is also capable of addressing both indoor and outdoor mapping.

As part of the Isaac AMR platform, NVIDIA DeepMap integrates with other components such as Metropolis, ReOpt, Isaac Sim via Omniverse and more.

NVIDIA Isaac AMR in action. The image is displayed as such because it shows a voxel grid created from point clouds and not the data from a camera, so it is sparse and is expected to be out of focus.

NVIDIA Metropolis adds real-time situational awareness

The mapping does not take into account everything in these environments. And the advanced sensors on board AMRs are not always enough to ensure safe and efficient operation.

The NVIDIA Metropolis video analytics platform meets the need for real-time external perception by providing access to cameras and sensors that are used throughout the factory or warehouse.

With Metropolis, AMRs have access to additional levels of situational awareness on the factory floor, enabling them to avoid heavily congested areas, eliminate blind spots and improve visibility of both people and other AMRs. In addition, the pre-trained models from Metropolis offer a head start when it comes to adapting to site-specific requirements.

ReOpt libraries are transforming logistics

NVIDIA ReOpt AI software libraries can be used to optimize vehicle route planning and logistics in real time that can be applied to AMR fleets.

There are many factors to consider when deciding the optimal AMR fleet size for use in large, complex environments. Robot speeds, battery life, transport size and weight, and facility layout all play a role.

Companies can (with Isaac Sim) simulate multiple AMR interactions with NVIDIA ReOpt. These can be done quickly and precisely in digital twins of environments like warehouses. And they can be implemented before robots are put into production when situations change, saving time and money.

And once they have been deployed, routes must be constantly re-optimized to achieve the greatest possible operational efficiency. NVIDIA ReOpt enables dynamic re-optimization of routes to a fleet of heterogeneous AMRs based on a number of constraints.

Provision of AMRs in production

The Isaac AMR platform, available on NVIDIA EGX servers, enhances the development of AI-driven logistics by providing a complete way to create industrial and human-robot simulations and route optimization.


Comments are closed.