Q How did Movel AI come about?
A We both have strong interest and experience in robotics.
Haoyu received his PhD in computer science from the National University of Singapore and was working on artificial intelligence for robots. He has developed algorithms for self-driving golf carts and the prevention of aircraft collision.
I have been working on robotics for the past five years, graduating with a master's degree in embedded systems from the Nanyang Technological University School of Computer Science and Engineering.
I started a hardware robotics company, selling educational robots to universities.
We both joined Entrepreneur First in October last year. It is a United Kingdom-based company which launched its international branch in Singapore and partners Infocomm Investments, now SGInnovate. They focus on individuals and bring together skilled people to form teams.
We met in the programme and came together to solve a problem that robot manufacturers were having.
Q What problem is Movel AI trying to tackle?
A We are making the next-generation robotics software. Robotics has been around for some time but it has not been able to penetrate the commercial space, which is what we would love to be able to do.
One of the major issues is the ability of robots to navigate and understand the world around them. They don't have the intelligence to understand the world around them and navigation is one of the major issues causing this problem.
Previously, robotics was using traditional sensors like expensive lidars, which measure distances using lasers. For instance, the lidar on the Google self-driving car costs at least US$70,000 (S$95,000). If you want to go into the commercial space, lidars would be very expensive.
Secondly, robotics manufacturers were also dependent on physical markers, and robots could only run and complete tasks based on the markers. But if these markers are blocked, the robot cannot complete the task.
Q How is Movel AI's software different from what's on the market?
A Instead of lidars, we are using computer vision using cameras. We take the input from the camera, and get feature points from the existing environment. Based on that, the robot is able to tell where in the environment it is.
Through our software, the robot gets more information about its surroundings. This helps to make it less accident-prone as it has better semantic understanding, such that the robot can identify humans, chairs and so on. The robot would not be able to do so using just lidar.
It also reduces costs as lidar sensors cost more than $22,000 while a camera costs only several hundred dollars.
The software is in development but we are running some pilots. We are doing a pilot with a robotics company, Unitec Megatronics. It makes robots for restaurants, such as for collecting dishes.
Most companies may be developing some software internally, but there are no third parties doing the equivalent of what we are doing.
Q What is your business model and what are your major challenges?
A We have got $10,000 in revenue from Unitec, and we are closing a seed round of funding next month.
We have received at least $120,000 worth of funding.
As this is a relatively new technology, it's difficult to penetrate the market with a customer base because customers will want to test the product, so we have a longer sales cycle.
The advantage is that when we penetrate the market, it is easy to expand as we would have a proven track record.
The major challenge for improving the product is to make it more robust and adaptive so it can work in many environments.
Q What are your plans for expansion?
A Our vision is that robots will help humans to have a better life. It's not about replacing people but coordinating with people to do work faster and better.
As we are working on something very technical, we hope to file a patent to protect us from competitors. The Intellectual Property Office of Singapore has been helping us navigate the patent landscape and we are in the process of filing our patent.
We hope to expand in Asia, especially in China, as there are a lot more robotics manufacturers there. We will visit Beijing to look at the market.
Once we have tested our products here, we plan to sell and expand mainly in China.