Engineer Spotlight: NXROBO’s Dr. Tin Lun Lam on the Complexities of Designing Home RobotsMarch 06, 2017 by Steve Arar
NXROBO's Tin Lun Lam shares insights on designing home robots.
NXROBO CEO Dr. Tin Lun Lam spoke with AAC's Steve Arar on the difficulties of the design process, machine learning, and how to build an effective team of engineers. The Hong Kong-based NXROBO has recently debuted a robotic butler named BIG-i.
Home robots have been lurking on the horizon for years. BIG-i is NXROBO's most recent contribution to this industry, a home robot which can interact with family members and smart appliances.
The robot uses a custom operating system designed by NXROBO to analyze voice commands and visual cues. The robot, which currently only speaks English and Mandarin, is programmed through if-this-then-that style voice commands. For example, parents can define different conditional situations like “If you see Tommy grabbing food, remind him to wash his hands”. As a result, with one sentence, users can generate custom functionalities.
BIG-i utilizes a wide variety of sensors to achieve different goals. Some of these sensors, such as tilting sensors and obstacle sensors, are employed to prevent the robot from falling or colliding with the obstacles. Other sensors monitor environment parameters such as temperature, humidity, pressure, and light, and allow developers to generate custom functionalities.
To perform audio localization, BIG-i relies on a 360-degree microphone array. An RGB-D camera, which is placed inside BIG-i’s eye, allows object tracking and recognition as well as 3D-map construction. To manage smart appliances, BIG-i utilizes Wi-Fi, Bluetooth, and infrared modules. According to Dr. Lam, the robot's soft body required that all of the sensors be housed inside the eyeball of BIG-i.
Interview with Dr. Lam
In order to get further insight into home robots and the process of their design, AAC's Steve Arar corresponded with Dr. Lam, CEO of NXROBO. As a bit of background, Dr. Lam holds doctorates in Automation and Computer-Aided Engineering from the Chinese University of Hong Kong.
BIG-i's "eye". Image courtesy of NXROBO.
AAC: Products like BIG-i haven't quite hit the mainstream yet, so there are plenty of challenges to overcome and questions to answer when it comes to design. What was the most challenging part of designing a home robot?
Dr. Lam: The biggest challenge we face is taking both people’s imagination of robot and practical feasibility into account. Public impression of robots comes mainly from science fiction and movies, but for making a robot in reality, we need to take a lot more into consideration. Nowadays, in the robotics field, there has not been a widely-used home robot yet, so most of time we can’t take an instructive precedent as a reference to solve problems. While making a robot, we need to think of both real cases and potential problems that have not yet occurred. We have to feel our way forward.
AAC: You've designed BIG-i to understand natural language and to be programmable through voice commands with the format of “if-this-then-that”. Do you see it being possible to incorporate an AI system that allows robots to analyze sentences in any format and perform tasks based on its own learning?
Dr. Lam: That’s one direction that we are forwarding. But there are still many difficulties to overcome because of the huge difference between human language and machine language. Compare to the latter one, the human language is much more diverse and random.
BIG-i in a home setting. Image courtesy of NXROBO.
AAC: In your Kickstarter video, you use "Hey BIG-i, if you hear a strange noise, please go check it out" as an example application of command recognition. Considering that there are so many possible interpretations of that language, how does the robot define a "strange noise"?
Dr. Lam: Now we have a sound library to help define a “strange noise”. If the sound is at a particular frequency or is louder than a certain number of decibels, the robot will classify the sound as “strange”. The database can be updated, so the definition will get more and more perfect in future.
AAC: What advice do you have for engineers designing products of their own, especially those pursuing crowdfunding for their designs?
Dr. Lam: You have to find the right partner—there’s nothing more important than this. It requires not only setting a common goal for all, but also finding partners that complement each other, both in terms of character and skills. You can create the right product only when you build the right team.
In terms of crowdfunding, you have to be clear on the reason you launched the campaign. Don’t be distracted. Focus on your original purpose will more likely lead you to success. You can get help from users with open communication.
AAC: BIG-i is a complex product. What was your team like and how did you divide the prototyping procedure into manageable sub-projects?
Dr. Lam: An intelligent robot is one kind of interdisciplinary, highly integrated product. It is impossible to complete the development process while only mastering one branch of knowledge, so the R&D team will need experienced talents in every specific technical field. Meanwhile, a responsible person is needed for managing work allocation, scheme determination, time distribution, etc. He needs to have rich experience in every related field to make the project more likely to succeed. Thanks to these kinds of talents in our team, we were able to develop the product in an efficient way.
The spec list of BIG-i—each entry on this list represents a different design challenge for Dr. Lam's team. Image courtesy of Kickstarter.
AAC: How long do you think will it take to see robots which can mimic human behavior such as pouring a cup of tea?
Dr. Lam: In certain environments, the technology is already mature enough for a robot to pour a cup of tea and do suchlike. Actually, industrial robots have already completed tasks much more complex and accurate than that.
For robots that serve at home, they have to face two main problems: If the robot is able to serve tea, how do we make it compatible with cups of different materials and different shapes? Meanwhile, if the robot can serve tea, that means it has certain athletic ability. We have to consider if this ability can lead to some unexpected consequences.
AAC: What additional functionalities would you like to integrate into future iterations of BIG-i?
Dr. Lam: Now we are making BIG-i learn new languages on some cooperative partners’ request. In the future, BIG-i will speak more languages so it will be able to communicate with more users from different countries.
Moreover, the NXROBO operating system BIG-i uses allows developers to develop applications freely, and these applications can create a variety of functions. For example, BIG-i can translate sign language and help deaf-mute people to communicate. It can also help to transmit holographic images so you can visit a museum while staying at home.
Meanwhile, we will corporate with more third-parties to add features such as online shopping and flight ticket or hotel reservations.
Thank you, Dr. Lam, for your time and expertise!
Featured image used courtesy of NXROBO.