March 01, 2017      

After obtaining more than $1 million in grants from the U.S. Department of Defense and the National Science Foundation, BluHaptics Inc. has obtained $1.36 million in Series A funding for subsea robotics. The Seattle Angel Fund also provided insight into what investors are looking for in startups.

Seattle-based BluHaptics was spun out of the University of Washington and has developed systems that combine machine learning, real-time 3D modeling, and haptic feedback to aid in the control of subsea robotics.

“The University of Washington is one of the largest public recipients in the U.S. of government research grants,” noted Susan Preston, a managing member of the Seattle Angel Fund. “Our fund is looking for any great deal in the Pacific Northwest.”

“We came out of the applied physics lab in 2013,” said Don Pickering, CEO of BluHaptics. “The initial problem was helping the Navy with subsea cleanup.”

“Such tasks used to involve human divers, then ROVs [remotely operated vehicles],” he told Robotics Business Review. “Even though they’ve been cleaning and repairing installations for 20 years, ROVs are still manually controlled and are not precise or efficient enough.”

“Things get challenging very quickly in dynamic environments,” said Pickering. “Software helps close the loop. We’re building software that enables safer and easier teleoperation of robots.”

Worldwide spending on robotics and related services will grow from $71 billion to $135.4 billion in 2019, predicted research firm International Data Corp. (IDC). After slowdowns in undersea exploration because of recession and low oil prices, that portion of the market is rebounding.

Business Takeaways:

BluHaptics has been developing software that combines 3D modeling, haptics, and machine learning to reduce the cost and training overhead of operating subsea robotics.

The Seattle Angel Fund, which led Series A funding for BluHaptics, looked for a unique product, a strong team, and a business model oriented toward commercialization.

The technology that BluHaptics is testing could have a wide range of applications, including more precise controls for robots and drones on land, in the air, and in space. However, humans are still a key part of the equation.

Dexterous ROVs an ‘aspirin, not a vitamin’

BluHaptics plans to use its funding to commercialize its software. The company is initially focusing on subsea robotics inspections and maintenance for the oil and gas industry.

ROV pilots need relief from the cognitive burden of keeping track of many variables. For instance, there are more than 600,000 connectors underwater, but they often break from torque from manual removal.

“The average cost of breaking an undersea connector is one-half to 2 million dollars,” Pickering said. “The goal is to reduce accidents and collisions while also decreasing training time.”

“With our software, the robot arm can move straight back. That alone is a significant advantage,” he explained.

“Any time we’re evaluating a company, one of the primary things we’re looking for is whether it addresses an actual market pain,” Preston said. “It has to address that pain with an aspirin rather than a vitamin, with an elegant solution, at a sound price point and format. A startup needs a business model that’s interesting to prospective customers.”

“We process data from lasers and sonar that normally takes four hours of processing work,” said Pickering. “The higher the resolution, the better for us. We’re using parallel processing and GPUs [graphics processing units] to make that data available.”

“Using 3D data, real-time rendering, and haptic feedback for avoidance, we can create ‘virtual force fields’ to prevent people from steering into objects or guide them with the exact approach every time,” he noted. “We’ll be conducting testing this summer, and we expect 25 percent gains in efficiency.”

BluHaptics has two patents, which helped it get funding. “The technology itself has to have some unique attributes and protection,” Preston said. “If the technology is not reverse-engineerable, that gives a company a distinct advantage in the market.”

“We’re the only software company tying together modeling and controls into a unified platform,” Pickering claimed. “With machine learning, this allows less-experienced pilots to have refined control.”

Software democratizes subsea robotics use

“There’s chronic problem in the subsea robotics world with underskilled pilots,” Pickering said. “Companies are interested in our software [so] they can take junior pilots and more predictably complete complex tasks.”

BluHaptics’ controls are based on technology from surgical robots and video games and can “deliver a level of precision that was otherwise not possible,” he said.

“The first product we’re releasing is a control upgrade for most widely used underwater hydraulic manipulator,” Pickering said. “It’s a $25,000 manipulator that’s hard to use with a seven-axis arm, but we enable it to be used with an Xbox controller.”

“Existing controllers can take up to six weeks to replace or get them shipped long-distance,” he said. “Having a $40 game controller you can buy at a grocery store is really enticing.”

“For younger users, that’s a much easier controller to train on,” Pickering noted. “We’re building a platform that is user-focused. You don’t need Ph.D.s to set it up.”

“Operators control only the gripper; our algorithm controls the joint and locks axes and maintains planes,” he said.

Teamwork leads to investment

Pickering’s background in software development and a diverse team not only helped BluHaptics develop its software; they also helped the startup get investment.

“The robotics space is very hardware-centric, and the vendors often are as well,” he said. “Yet I would argue that software is the only way to move beyond purely manual control in teleoperation of robots. Sensors getting better all the time, but you need parallel processing and machine learning to deliver technology in a way that’s intuitive to the operator.”

One reason investors may be reluctant to fund robotics is that hardware has more capital requirements, said Preston. “The overall financial model is that they can scale their returns a bit better from software,” she said. “Still, we support some, such as medical devices.”

Subsea robotics executive Don Pickering

CEO Don Pickering and the BluHaptics team have diverse software and technology experience.

“I was a vice president of business development at Microsoft research, and our staff is a combination of software engineers, physicists, and electrical engineers,” Pickering said. “We recognize the opportunities and arc of technology development.”

“We look at the team. The No. 1 reason companies fail is bad market timing, but the No. 2 is bad management,” said Seattle Angel Fund’s Preston. “You can have great technology and address market pain, but you need a team to execute on plans. … We really like BluHaptics’ team, which is conscientious, responsive, and has early traction with clients. In our reference calls, we heard all the right things.”

Interoperability and other applications

“We’ve developed our technology to be sensor- and platform-agnostic,” Pickering said. “If it’s 3D data, we’re working with point clouds. It can work with sonar and stereoscopic cameras.”

“This will revolutionize subsea robotics,” he added. “Underwater robots can cost $5 million, $50,000 to operate, and millions to replace. We’re getting a lot of traction and interest. We’ve done our first integration tests with an oil and gas operator and have more in a few weeks.”

BluHaptics’ software could also improve control of robotic arms in terrestrial and space applications.

“We’ve talked with customers in industrial automation,” he said. “They use 2-D systems because they can’t overcome latency for 3D. We unlock that capacity. In manufacturing, our software enables customers to work with large dynamic objects like aircraft.”

“We’ve also had some interest in our technologies for AGVs [automated guided vehicles] for route optimization and to prevent rollover,” Pickering said. “There are certainly analogues with self-driving cars, where we started with cruise control, parking assist, and lane notification. There has been an assumption of assisting drivers or relieving them of tasks one by one, and to be able to respond to the unforeseen.”

Still, any robotics investment is a bit of a gamble. “We saw about 90 companies, took 15 through diligence, and made six investments,” Preston said. “When I teach angel investing, I say that seven out of 10 won’t return capital but might return a portion. That’s one reason we argue for a lot of diversification with 10 to 20 companies in a portfolio; then you have a higher chance of choosing winners.”

“BluHaptics has been in discussions for IED [improvised explosive device] and mine defense, as well as with space organizations, because their platforms are video-only and do not leverage 3D data in real time,” Pickering said. “We’ve even had interest from a major aerospace company, which is considering putting manipulators on drones inspecting dams.”

“It’s possible to use our technology for dynamic positioning and stabilization for drones — it’s still a little ways out, but there are opportunities,” he said. “Eighty to 90 percent of our code is transferable across all these fields; we just need to develop drivers and connections to existing controls. There are still massive applications beyond self-driving cars, and this could eventually filter down to robots in homes.”

“Startup founders need to understand the long-term goal of liquidity. They need good partnerships and should have the goal of commercialization now rather than in three years,” Preston said. “It’s also important to have a market that can scale and have an exit strategy. An acquisition might not be worth $1 billion, but you may need only $10 million to get a good return.”

More on Investments and Subsea Robotics:

Robots need humans

“We’re starting to automate routine responses to standardized objects,” said Pickering. “Ultimately, we’ll start to see robotic intervention, repair, and installation take over more tasks.”

However, humans are still an essential component of robot operations. “Look at the DARPA Robotics Challenge — the teams that consistently won had a human-robot approach,” Pickering observed. “Those that focused just on automation failed.”

“Having a human in the loop is not negotiable, certainly with high-risk operations undersea or in space,” he added. “Using technology to assist humans is the first step.”