How Levatas Teaches Spot New Tricks
One of the key components to Levatas’ success: partnerships. Customers looking for accurate analog gauge reading and thermal heat detection look to Levatas; in turn, Levatas partners with other AI industry experts to maximize operational efficiency, product innovation, and every AI opportunity available.
Case in point: Spot, the amazing robotic dog created by Boston Dynamics, is trained by Levatas to be a necessary workplace asset, processing visual data in hundreds of different use cases. Levatas looked to Boston Dynamics because the company is a pioneer in mobile robotics, combining sophisticated mechanical designs and state-of-the-art electronics and software.
“We don’t want to be robotics creators,” Daniel says. “We don’t want to reinvent the wheel on Enterprise software. So we pull in partners who really specialize in those areas. We’ve been early adopter partners with Boston Dynamics for about eighteen months now. We were one of the first companies to build payloads [transmitted data], especially computer vision payloads, for Spot. We are working with best-of-breed partners right now. There is no close second to Boston Dynamics. That’s why the strategy of working with partners makes so much sense.”
Spot is an agile mobile robot that maneuvers through terrain with ease and precision, allowing companies to automate inspection tasks and capture data accurately. It makes routine operations safer and more predictable. Spot allows outside hardware to be attached and integrated using mounting rails and payload ports; pre-built solutions from an existing network of third-party software hardware providers are easily utilized. As a result, custom controls can be created and autonomous missions can be programmed. If that wasn’t enough, here’s a huge selling point: the ability to integrate sensor inputs into data analysis tools. Spot has become a favorite of many industries, including construction, healthcare, entertainment, mining and public safety.
For Levatas, it’s not just about merely selling a cool toy. If you just have a robot that walks around, you’re not getting any value from it. Instead, Spot is designed to capture image data, and lots of it. It has a sensing capability that lets clients not just capture data, but gain insights from it.
Levatas’ work with Spot brings opportunity to clients, automating tasks that were originally labor intensive, such as:
- reading analog gauges
- performing facility inspections
- detecting anomalies on a production line
- monitoring equipment
- protecting employees and improving safety
Spot uses its robotic arm and an internal AI processing chip connected to a thermal camera to integrate sensors or custom payloads based on use cases. If you have hard-to-navigate terrain that’s not particularly people friendly, Spot will create missions in order to autonomously navigate your planned routes. It can also use AI models to determine route calculations and movements. Levatas’ Vinsa technology will help image and video data train models over time (with human help, of course), so that Spot will learn and increase accuracy over time with consistent feedback.
The result: immediate reductions in operational costs.
In many cases, clients come to Levatas with brand-new cases that nobody had ever solved before. The end goal: to build an integrated end-to-end solution.
Daniel says, “The story we’re telling is a blending together of state-of-the-art robotics as a capture platform. It’s the idea of putting together information at your fingertips that you’ve never had available to you before at a pace that you’ve never experienced before. We marry these together and add in a uniquely human factor that brings value and solutions.”
Daniel, who has been with Levatas for fifteen years (the reason: “we feel like we have something special going on.”), regularly connects business problems with Computer Vision AI solutions. After vigorous data aggregation and prep, a champion model is selected, then deployed and monitored. Feedback loops are created to ensure that models don’t ever become stale in production. The goal is to help computers sense the world around them in a way humans are able to do.
One of his job’s biggest rewards: helping customers understand this amazing technology and how operationalizing AI solutions will deliver value.
Daniel says, “the idea is that you want customers to have faith in you and trust you to know that you’re getting them where they need to be.”
What are clients typically asking for? The bottom line: helping them to figure out how to use data to make smart decisions.
“These days, most of our customers are trying to draw insights out of visual data,” Daniel says. “For instance, not just looking at a pipe, but a pipe that’s leaking, or a gauge that’s reading at 90 PSI, meaning we have to take some kind of action because equipment is malfunctioning.”
With that in mind, consider another Levatas client category: the entertainment and tourism space. Levatas uses visual data to recommend improved guests experiences; for instance, a cruise ship restaurant with a long line could lead to improved traffic patterns and flow.
The near future is also bringing exciting innovation from Levatas in the form of combining robotics, like infusing robotics and drones with new computer vision solutions to further improve gauge readings and inspections. Imagine Spot sniffing around a nuclear power plant looking for anomalies, so that humans don’t have to.
“We’re here to increase our clients’ quality of life,” Daniel says. “Let’s take away the jobs that most people don’t want to be doing anyway so that they can focus on things that require a uniquely human intelligence package. It’s never about taking humans out of the loop to be replaced by AI. We like to say we’re in the business of augmenting humans, not replacing them.”
Find out more about our content partner here.