|
Formant |
|
|
November 4, 2019 Issue |
||
|
CEOCFO MAGAZINE |
||
|
|
||
|
IFormant is Enabling Robotics Companies to Observe, Analyze, and Operate Robot Fleets |
||
|
Chief Executive Officer
Formant www.formant.io
Contact: Jeff Linnell Jeff@formant.io
Interview conducted by: Lynn Fosse, Senior Editor, CEOCFO Magazine, Published – November 4, 2019
CEOCFO: Mr. Linnell, what is the concept behind Formant? Mr. Linnell: We enable businesses to automatically collect, analyze, and act on robotics data. Specifically, we enable operators to keep robots active in the world.
The thesis is that this next decade of automation is going to be empowered by humans assisting machines as they venture out of the factory and into our world. The big shiny object there that the world can relate to is self-driving cars. It is going to be a very long time before we take a ride in a self-driving car. That said, there are steps between here and there, like robots working on construction sites digging foundations. That application is absolutely doable and it is going to be made possible through human assistance. For example, a robot with an issue will have an operator that can facilitate that robot’s mission for next decade of automation. Robots leaning on human beings to get their work done.
CEOCFO: How does the Formant platform work? Mr. Linnell: Everything that we do is cloud and browser based. You install a piece of software on your robot that is listening to its signals and continually transmitting them to the cloud. These data streams include content such as video, lidar, 3D scans, and telemetry. It might be the torque values from a motor, the amperage of a battery, its location in the universe and all its log files. You are continually getting all this information off of a robot and that is presented to a human operator so they can monitor in real time the status of a robot, anywhere in the world.
The operators will see charts going by, live streaming video, or more likely they’ll see alerts when robots hit some sort of an obstacle to doing their job. A robot might, for example, be unsure about what they are seeing in their camera so they need instruction to go left or right, or they have hit an obstruction and want to know if it safe to proceed. If I broke down the system for you, part of it is on robot software that is listening, compressing and transmitting that information to the cloud and once it is in the cloud, it is observable through a web interface.
As you go down the chain, there are other services on top of that. When your data is flowing through the platform, we make it more intelligent so that you can perform anomaly detection and analytics on your data streams or search a video image stream semantically. We provide data ingestion, presentation and enrichment.
CEOCFO: For Formant purposes, what is a robot? Mr. Linnell: Our interpretation of robots is extremely broad. We are particularly interested in robots that have a job to do that might be just possible, or working 95% of the time, and could use a (human) hand. We are working with drone companies, industrial floor cleaners, excavators, robots assisting trades in construction, and even robots performing animal behavior training. The commonality is a certain amount of sensing, some actuation, and certain amount of variability in the task or environment. More often than not these machines resemble purpose built tools rather than the canonical “Robot” or “Robot Arm” that people tend to imagine.
CEOCFO: Are companies that use robots or developers of robots understanding the need for more and how are you helping them do so? Mr. Linnell: Today, the majority of robots that are in the universe in significant numbers are welding robots and Roombas. Otherwise there are research projects and million dollar machines that do not really perform a task. In my opinion there have been a lot of failed applications historically due to the inability of a robotic system, or a misjudgment of a tasks utility.
This is changing. People are becoming really pragmatic now and they are taking rifle shots at specific applications. They are not even thinking about these systems as robots, but simply as adding automation to a task. For example, you might find a robot applied to a relatively tractable problem like digging a foundation on a construction site, or mopping a floor in a building. Both of these are happening, and relatively constrained problems. They are also large commercial opportunities. In the case of floor cleaning, the robot is simply thought of as a way to move the mop. The business delivers clean floors as a service. The robot is irrelevant to the end customer; in fact these companies often refer to the machine as a “tool” as opposed to a robot. It has a certain amount of autonomy, but is often assisted by a human co-worker to handle the edge cases, like corners or removing an obstruction.
The combination of these pragmatic applications, cheap sensing, and modular mechanical components has led to robots being deployed in fleets numbered in thousands with aggressive scaling anticipated. The scale of these deployments and value of the application data has created the need for fleet management and operations software. Most companies would prefer to leverage a scalable built platform that can accelerate their time to market and assist in the ‘care and feeding’ of their robot fleet as opposed to reinventing the wheel. Formant’s platform assists these companies in three ways. First, we make it easy to reliably gather robot data from field deployed systems. Secondly, we make it easy to introspect the robots health and performance at the individual system level, or fleet wide view. And finally we allow humans to command the robots from a centralized interface. This could take the form of literally 'beaming in’ to a specific robot via a web page, or larger gestures such as sending an updated algorithm, or software patch to thousands of robots at a time using our command line interface.
The robotics industry is starting to wake up to the fact that this tooling is required to manage systems at scale. If a company has been around for 5-10 years, they have likely built significant portions of this infrastructure, however the technology choices might be antiquated. If a company is looking to stand up a RAAS (Robots As A Service) business, they are eager to get to market quickly and are appreciative for this tooling.
CEOCFO: How do you reach potential customers and where should you focus to start? Mr. Linnell: We’re focused on US-based robotics companies. Because we’ve been in robotics for so long, we’ve got a pretty good sense of who the major and minor players are and keep track of who is getting funding and expanding. We’re relentlessly customer-focused and having as many conversations as possible with leaders in these companies. After a handful of these conversations, we discovered that there is a need for tooling in the observability space. There is a lack of human-usable interfaces to analyze and observe robot fleets.
Focus; however, is my largest challenge. Deep down, I want to do everything, but it is important to find the white space and focus on problems the market is asking for now. The key is being very aware of the competition and listening to customers to find white space. Right now, the market is small, but this is very quickly going to become too big for any one company to dominate, so you are going to have expertise in certain domains.
CEOCFO: Are you funded for your next steps or seeking investment or partnership? Mr. Linnell: We have been well funded to date. When we started this, it was not difficult to raise money. I attribute this to the pedigree and expertise of our team, and general awareness that software is the next frontier in robotics. The venture community is using the language of our industry. Automation is clearly happening, and we are well positioned at the intersection of human involved robotics and cloud data.
CEOCFO: What surprised you as Formant has grown and evolved? Mr. Linnell: What surprised me was how quickly the industry is waking up. I felt like when we started this, we were early. But now, I think we were at exactly the right place at the right time. There is a healthy amount of competition and people are getting really pragmatic about robotics. It is starting to scale. Robotics over the last twenty years had many fits and starts, where people thought if they could just solve one problem - the manipulation or actuation - it would take off.
The trend I am seeing now is that companies are getting real work done with these machines and showing that a human plus a number of these systems can be twenty times as effective as either of them individually. Where we are in San Francisco I could point to seven robotic companies within three blocks of us here that are automating very specific tasks with machines. To me, that is a giant signal that this is finally starting to get some traction. The robot future will be different than we thought it was going to look. It is not the butler and the burger flipping robot, it is a machine that is doing some work and likely with human involvement.
|
“The robotics industry is starting to wake up to the fact that this tooling is required to manage systems at scale. If a company has been around for 5-10 years, they have likely built significant portions of this infrastructure, however the technology choices might be antiquated. If a company is looking to stand up a RAAS (Robots As A Service) business, they are eager to get to market quickly and are appreciative for this tooling.”- Jeff Linnell
|
|
|
disclaimers |
||
|
|
||
|
|
||
|
|
||
Jeff
Linnell