Building robots to better understand human behavior is an intriguing idea. This is Hiroshi Ishiguro’s unusual approach to robotics. Ishiguro has been an engineering and robotics pioneer ever since is Intelligent Robotics Laboratory took off in 2004. Here is what you need to know about this forward-thinking scientist and his creations.
Who Is Hiroshi Ishiguro?
While most scientists design robots with specific tasks such as climbing obstacles, working in factories, or even delivering packages in mind, Ishiguro focuses on the relations between humans and machines. His creations stand out with their lifelike designs.
Ishiguro is mostly known for creating realistic androids that can mimic human expressions. These androids can be controlled remotely by an operator who wishes to project his or her presence in another location.
Ishiguro was originally going to become a painter. He changed his mind after taking a few computer science classes.
He graduated from the University of Yamanashi in 1986 and went on to earn a grade degree in Engineering Science from Osaka University. He started teaching shortly after graduating in 1991. His teaching resume includes positions at prestigious Japanese universities, such as Kyoto University, Wakayama University, and Osaka University.
Ishiguro eventually retired from teaching to focus on building robots. He created the Intelligent Robotics Laboratory and founded the Hiroshi Ishiguro Laboratories a few years later.
He has won several awards such as the Best Humanoid Award at the 2006 RoboCup held in Bremen, Germany. He is also well-known for being the co-author of Human-Robot Interaction In Social Robotics.
This concept of social robotics is at the core of the projects currently in development at the Hiroshi Ishiguro Laboratories. Ishiguro is overseeing a team of researchers who work on studying interactions between robots and humans. The team has developed humanoids that can replicate a human presence and is also creating a technology that would allow humans to transmit their presence via a cell phone.
Why Is Ishiguro’s Approach Different?
Robotic scientists often design machines for a specific task. The design of the robot is strongly influenced by the environment in which the robot will be performing the task and by the task itself. For instance, factory robots are often designed as a single robotic arm mounted on a platform.
Ishiguro focuses on how humans will respond to the robots. He often uses human models for his humanoids. He has created an android robot modeled after his image, the Geminoid HI-2.
While scientists typically have a background in robotics and engineering, Ishiguro uses an approach that is also grounded in neuroscience and cognitive science. The purpose is to make people feel as if they were in the presence of another human rather than creating a machine that can perform a task.
Meet Ishiguro’s Creations
Hiroshi Ishiguro and his team of scientists have created several robots over the year. These are the ones that stood out the most thanks to their lifelike appearance.
The Geminoids are a series of lifelike androids based on Ishiguro himself. You can see the Geminoid HI-2 in this video.
Ishiguro recently completed the Geminoid HI-5. Each version of the Geminoid is more advanced and more lifelike than the previous one.
The development of these androids has helped Ishiguro and his team progress with the use of actuators to mimic human movements. The Geminoid HI-2 human android had 13 actuators in its face, 15 in its torso, and 22 in its arms and legs.
These actuators allow the android to replicate a wide range of movements and facial expressions that one would be able to observe when interacting with another human. The Geminioid HI-5 even has a vision system that captures how the operator’s lips move to reproduce these exact movements.
A lot of people feel a little uneasy when facing one of these androids for the first time, but the researchers report that a lot of people start to relax after a while and will engage in a conversation with the Geminoids.
There have been two versions of the Repliee android so far, Q1 and Q2. This small female robot is based on Ishiguro’s daughter.
There were a few issues with the Repliee Q1 model that caused people to feel very uneasy. The robot’s facial expressions were very stiff, and the android would shake when performing some movements.
This robot helped Ishiguro and his team study all the small movements that a woman would perform during social interactions, such as blinking and hand movements. All these small details were captured and written into the Repliee Q2 software to create a more natural presence.
The Repliee Q2 android is one of Ishiguro’s most impressive creations thanks to actuators that reproduce breathing and small head twitches.
While the Geminoid and Repliee models are meant to be telepresence robots, Erica is a project that aims at creating a robot people can have a conversation with.
Erica is equipped with features like voice recognition and human tracking. The robot’s software was created to generate natural motion. She speaks with a synthesized voice rather than letting an operator speak through her.
The purpose of the Geminoid F was to create a teleoperated android on a smaller budget. This is one of the keys to making this technology accessible to the public.
The software developed for the Geminioid HI-2 was used for the Geminoid F as well. However, this android was equipped with only 12 actuators.
The Geminoid F was used to project the presence of an actor during a stage play in 2010. The actors who interacted with the android said the experiment was fun but they didn’t feel a real presence on stage.
Other Telepresence Projects
Not all the telepresence robots created by the Ishiguro’s laboratories are lifelike. Some of the telepresence androids have a very basic design. These projects include the telenoid, elfoid, and hugvie robots.
This is the kind of technology that could become mainstream in the future since these androids will be a lot more affordable and can be operated by a cell phone. Everyone could use these androids as a telepresence system rather than using a lifelike robot that mirrors their appearance.
What’s Next For Ishiguro?
Hiroshi Ishiguro has stated that his goal is to figure out all the elements that are unique to a human presence and build them into a machine. He would like to be able to give intentions and desires to robots since this would help the robots better understand the humans they interact with.
The development of AI and deep learning algorithms has allowed Ishiguro to make some amazing progress over the past few years. Deep learning algorithms have helped him develop better interfaces for the telepresence androids and could result in robots eventually being able to automatically reproduce human expressions.
In a way, human emotions and reactions are programmed with hormones and other chemicals. Ishiguro believes it is possible to create robots that could be programmed to feel and react just like humans.
Will Ishiguro’s Creations Ever Become Mainstream?
The androids created by Ishiguro and his team look like they walked out of a science fiction movie. His quest to create the sensation of being faced by an actual human could have a significant impact on how we use telepresence systems.
In theory, it is possible to trick the human brain into thinking that an android is an actual human. However, the android’s movements and facial expressions would have to flawlessly mimic human emotions and reactions. If this could be achieved, your brain could believe you are faced with a real human and even become attached to this person.
However, this would be very difficult to achieve. The slightest mistake in how emotions are transcribed or even a delay in showing a reaction will result in a human feeling extremely uneasy. This negative reaction can be very strong and border on terror, especially if the android’s facial expressions don’t match its behavior.
This phenomenon is known as the uncanny valley. This is why most people are initially uneasy when finding themselves face to face with one of Ishiguro’s creations.
This can happen with lifelike dolls, androids, and even CGI characters in movies and video games. If something looks close enough to a human, your brain will automatically start looking for subtle cues such as facial expressions that indicate how the person feels. You will start feeling uneasy when your brain can’t pick up on these small cues.
The uncanny valley phenomenon is a crucial obstacle for Ishiguro and his team. His creations would have to be impossible to distinguish from an actual human to avoid triggering these negative reactions. Some scientists believe that we might become even more proficient at looking for these subtle cues if lifelike androids become a common occurrence.
New technologies like AI and deep learning have opened new doors for Hiroshi Ishiguro and his androids. The next generation of robots will be even more lifelike than the current one. However, the difficulty might not be in creating more performing technologies but in figuring out all the subtle cues that indicate a real human presence.