Realtime Robotics

[MUSIC PLAYING]

PETER HOWARD: So my name is Peter Howard. I'm CEO of Realtime Robotics, where we are working to enable a new wave of robotic automation with breakthrough innovations that make creation and deployment of robotic systems many times faster, much less expensive, and more effective.

There's two separate status quos, if you will. There's the status quo for the industrial robotics-- traditional industrial robotics-- which is very painstaking programming of robotic systems, point-by-point, step-by-step. And that's something which we have been able to completely automate.

So instead of a programmer going through and spending hours, days, weeks, months programming robots in a system, we have completely automated that. And that reduces the programming component of building a robotic system by upwards of 90%. We're what you might think of as the more modern robotic implementations, where people are using sensing systems and AI to identify objects and so on.

The unfortunate fact is that even if you've got these beautiful new AI systems sitting on the top, they still have to work with the same fundamental problems that underlie the industrial robotic systems, which is you still have to program motion. You still have to work with a variety of different interfaces that different robot makers have and the dynamics of each robot.

So the way we're addressing that is to basically solve those underlying problems of robotic interfaces, ELC interfaces. We solve those once, and we offer that solution as part of our platform.

That platform can be used by these new system makers, these new AI-based system makers, so that they can stay focused on what they are really good at, and not have to pay attention to all the difficult underside of the robotic problem.

First and foremost, our platform is based on autonomous robot motion planning, and multi-robot deconfliction. We achieve that through pre-computing of a field of potential motions that the robot is likely to need for a given application. And that can be tens of thousands, hundreds of thousands, tens of millions of motions, which are pre-computed.

And then we hardware accelerate the searching of those potential motions at runtime. And are able to do that with the hardware acceleration in such a way that we can look at all of the potential options, see moment-to-moment, millisecond-to-millisecond, which ones of those are available for use, and then find the optimal path through the workspace to get the job done.

So that's our core technology. So how do you get a single or multi-robot system to achieve the maximum throughput possible? We do that through an AI for multiple robot optimisation, which works for both highly-structured and semi-structured, unstructured work cells.

We're able, through running millions and millions of iterations, to find the best and highest efficiency structure for the work cell. So this includes everything from the positioning of the robot to the sequencing of the work, of the tasks which the robots are going to do, the allocation of which tasks are going to be done by which robot, et cetera, et cetera.

So that in the space of a few hours of running this AI, you're able to achieve a throughput rate that is just unimaginably better than what a human programmer is capable of doing.

Then finally, flexibility, which is the other Achilles tendon of the robotics world, is achieved through incorporation of powerful spatial and object-perception pipelines in the compute [INAUDIBLE] platform which we have created.

And that can be used for collision avoidance, it can be used for work piece perception, but identifying the object or what's happening with the transformation that you're doing, performing on that object.

But also, and extremely importantly is human safety. We will be putting on the market the first system that is capable of interacting intimately with people and keeping them safe, even in the presence of industrial robots.

So we look for those who are willing to put significant skin in the relationship early, so that if we meet our commitments, we know that they've already sold, all the way throughout their organization, the importance of the relationship.

Some of these relationships have taken two years to hit the market. The fastest is the most recent, and it's public now, which is Siemens, which only took seven months, start to finish, first introduction in the market.

The next six to twelve months are truly exciting times for us. With the global automation OEM leaders now promoting our products, the top 10 automakers doing the first rollouts of our products and incorporating them in their standard tools and workflows, with breakthrough new capabilities for optimization and safety being added to our platform, as well as tanking up a little bit on the upgrades inside, it's going to be a really exciting time.

[MUSIC PLAYING]