Mobile robot operations are becoming more sophisticated, taking advantage of the cloud and simulation to manage and scale more varied fleets. InOrbit Inc. today announced that it is working with NVIDIA Corp., Ekumen Labs, and Clearpath Robotics “to blend reality and simulation in robot deployments.”‘
“Our collaboration with Ekumen and NVIDIA shows for the first time how simulated and real robots from multiple vendors can be orchestrated through a single platform, allowing operators to switch seamlessly between them,” said Florian Pestoni, co-founder and CEO of InOrbit, in a press release.
“NVIDIA has built a lot of the infrastructure — not just hardware, but also software — for robotics simulation,” he told Mobile Robot Guide. “Being able to build on its technology has made a big difference.”
InOrbit is a member of the NVIDIA Inception program, which is designed to help startups by providing access to the latest technical resources.
“Ekumen is a custom software developer and strategic partner that has worked with a lot of well-known companies in the robotics space,” Pestoni added. “If you’re a large 3PL [third-party logistics provider] and have a massive warehouse, Ekumen will work with you to build a model of that warehouse and install virtual robots.”
InOrbit has also collaborated with Clearpath and OTTO Motors, which Rockwell Automation recently acquired, because “without robots, there’s nothing to simulate or manage,” said Pestoni. “They had already used simulation, but now we can support the next level of interoperability.”
“What we do at InOrbit is bring those pieces together under a ‘single pane of glass,'” he said. “It really takes a village as the industry matures.”
Mountain View, Calif.-based InOrbit said its multi-cloud platform can manage robot operations, or RobOps, at scale. The company said it provides observability through secure, real-time analytics and data collection, robot performance monitoring, incident response, and root-cause analysis.
InOrbit accurately simulates Robot Space
InOrbit said its partnerships will enable end users such as 3PLs to bridge “the gap between cutting-edge, AI-powered robotics technology and real-world applications.”
“We talk with some of the largest 3PLs out there, and they have autonomous mobile robots [AMRs] from a dozen vendors at their sites around the world,” noted Pestoni. “While they could give each its own space, that doesn’t really scale. As you grow to hundreds or thousands of warehouses, each one is different. We’re also seeing more specialized robots for tasks like trailer unloading, picking, inventory, cleaning, or pallet moving.”
“Once you get a taste of how to be more efficient for one task, you want to bring that to all tasks,” he said. “We see our role as helping to scale robot operations across vendors, across sites, and across the world.”
Users can define the recently launched InOrbit missions, execute the missions in both simulated and real environments, and then analyze the results, according to the company.
InOrbit explained that it uses the NVIDIA Isaac Sim powered by NVIDIA Omniverse to model complex environments and robot behaviors. The company worked with Ekumen on a demonstration (see video below).
The demo shows a virtual Robot Space that mirrors InOrbit’s actual Robot Space in Silicon Valley. Viewers can see multiple simulated robots performing the same missions as their physical counterparts.
“Whether it’s picking items at a fulfillment center or transporting boxes in a warehouse, the simulation provides a realistic representation of the physical world,” said the company, which offers tours of its virtual Robot Space through team@inorbit.ai.
“We were able to model robots from multiple vendors working together,” Pestoni explained. “We’ve been doing that in the real world for some time, but the value of simulation is that customers can try many things and pick the one that works best.”
“The challenge for a lot of companies is that if you have to start over when you switch from simulation to reality, you may not get the same result,” he added. “Our platform defines behavior in a robot-agnostic way, so users can get the exact same analysis and missions running in both. They can then compare apples to apples.”
Real and simulated data helps anticipate issues
The combination of real and simulated data plus RobOps tools will enable 3PLs, manufacturers, and retailers to derive actionable insights, asserted InOrbit. Users can model changes to workflows, orchestrate robots from different vendors, and reduce the costly impact of traffic jams or collisions, it said.
InOrbit added that its InOrbit Connect certification program allows partners across industries to optimize productivity. It cited the example of OTTO Motors by Rockwell Automation, which can see a simulation of its robots in a virtual twin of a customer’s facility.
“By using a single pane of glass to bridge the Sim2Real (simulation to reality) gap, customers can anticipate problems before they occur and continuously improve their operations,” said InOrbit.
Simulations could facilitate mobile robot interoperability
“Simulations already play an important role in the development of robots, such as for use in reinforcement learning,” stated Ryan Gariepy, chief technology officer of Clearpath Robotics and OTTO Motors by Rockwell Automation.
“As part of our efforts to advance interoperability, we are glad to collaborate with InOrbit to show how robots from different vendors can be simulated to help end users visualize their robot operations in real time before they deploy in production,” he said.
“The notion of orchestrating heterogeneous environments with multiple robots was ahead of its time three years ago, but InOrbit now gets a constant inflow of customers that have one robot for x, another for y, and want to add a different robot for z,” noted Pestoni. “They may have worked fine on their own, but put them together, and not so much.”
“As the industry matures, the idea of one company building everything from beginning to end doesn’t make sense, but our partner ecosystem has brought together the processors, software, simulation, and orchestration for AMRs to work together,” he said. “Users can mix and match robots from multiple vendors using the same traffic-management logic. They can test it out virtually and with InOrbit and have greater assurance when they’re ready to deploy robots from their next vendors.”
