skip to main content

Humanoid Robot Brains

6 minutes

(Describer) Science Out Loud.

(male) When you jump, climb a tree, or play basketball, you're doing things that no robot can currently do. You're outperforming the work of thousands of scientists, companies, and even governments. For you and me, moving around is easy. But for a robot to control its motion, it's much harder. If we could copy our motor control in robots, we could replace humans on hazardous jobs like firefighting, performing a spacewalk, or cleaning up the meltdown at the Fukushima nuclear reactor.

(Describer) He steps to a power drill.

If I want this drill, I pick it up. There's no thinking involved. But why is that?

(Describer) He squeezes the trigger.

It's because my subconscious mind recruits billions of brain cells to do the complex calculations involved for me. When I decide to pick up the drill, 30%-50% of my brain subconsciously processes the signals from my eyes.

(Describer) Other objects are around.

Now, in the front part of my brain, I consciously plan to reach my arm out towards the drill, grasp it, and bring it back to my body. My brain subconsciously transforms these vague commands, first, to basic motor plans in my premotor cortex, then, more precise muscular instructions in my motor cortex. Finally, my cerebellum perfectly coordinates activation patterns that get passed to the brain stem and sent to all of my muscles. My brain is constantly comparing the plan with the real-time situation and adapting to what I'm sensing. You see why it's so hard to get robots to do complex tasks. It's not just controlling a motor. You have to plan and perfectly adjust all of the motors together.

(Describer) A toy monster truck speeds over a floor.

With robots that have wheels or treads, we don't worry about balance. We always have a stable base.

(Describer) A rectangle for the truck.

But if a robot encounters stairs or a ladder, its mission could end suddenly. That's why we're developing bipedal robots that have two legs. Though they're harder to control, they're more adaptable and better suited to work in a human environment. This bipedal Atlas robot, made by Boston Dynamics, is being used by the M.I.T. team to compete in the 2013 DARPA Robotics Challenge. Call him Helios.

(Describer) Taller than the host.

M.I.T. is working hard to develop Helios' brain, which is a series of algorithms, or steps that Helios can take to take his sensor data to control his 36 joints and motors. Without these algorithms in place, Helios is really just a very expensive, 330-pound pile of metal.

(Describer) He knocks the chestplate. A man works at a bank of monitors.

The M.I.T. team is working hard on algorithms that improve Helios' ability to walk on difficult terrain.

(Describer) Helios steps on concrete blocks.

Currently, a human operator must send commands to Helios through a wirelessly connected computer. The operator looks at Helios' sensor readouts and decides what basic actions the robot should do. It's like how my frontal lobe took visual input from the back of my brain and made plans for my motor system to execute. Helios' operator makes these decisions and sends commands to Helios, who executes them himself. Helios' brain is kind of a hybrid. It's part human and part computer.

(Describer) ...the operator.

(Describer) Helios.

But why do we need our human operator? That's because as difficult as motor control is, motor planning is even harder. It takes vast knowledge about your environment to translate your goals into plans.

(Describer) Helios’ arm rises.

To us humans, it's obvious how to pick up this drill because we know so much about it. We know that it's solid, it's held up by the table, and it's picked up from the handle. But to program all of this information into Helios would be impossible to do for every conceivable object. However, some labs are working on other robots that can learn this information slowly from its environment, like we do when we're infants. For Helios to complete a task on his own, his creators must figure out how to adapt to an ever-changing environment. If someone takes this drill before I pick it up, I can run after them and take it back.

(Describer) He squeezes the trigger and sets it down.

But if I take this drill from Helios, he could not perform without his human operator. It takes 30%-50% of our brainpower just to process all the sensory information from our eyes. So you can imagine the challenge of programming a robot to continuously pay attention to his surroundings and adapt to them.

(Describer) Helios picks up the drill.

Fully autonomous robots are on their way. But partially autonomous robots, like Helios, can still be helpful while we're waiting. Someday, firefighters will control robots that can carry people from burning buildings, and astronauts might do spacewalks remotely, controlling robots like Helios from Earth.

(Describer) Title: Made with love at MIT. Accessibility provided by the U.S. Department of Education.

Accessibility provided by the U.S. Department of Education.

Transcript Options


Now Playing As: Captioned (English) (change)

Report a Problem

The smartest people in the world have spent millions of dollars trying to develop high-tech robots. Even though technology has come a long way, these humanoid robots are nowhere close to having the "brain" and motor control of a human. Why is that? A MIT scientist explains the motor control processes in the human brain, and how cutting-edge research is trying to implement it in robots. Part of the "Science Out Loud" series.

Media Details

Runtime: 6 minutes

Science Out Loud
Episode 1
4 minutes
Grade Level: 9 - 12
Science Out Loud
Episode 2
6 minutes
Grade Level: 10 - 12
Science Out Loud
Episode 3
5 minutes
Grade Level: 10 - 12
Science Out Loud
Episode 4
4 minutes
Grade Level: 10 - 12
Science Out Loud
Episode 5
4 minutes
Grade Level: 8 - 12
Science Out Loud
Episode 6
6 minutes
Grade Level: 9 - 12
Science Out Loud
Episode 7
6 minutes
Grade Level: 10 - 12
Science Out Loud
Episode 8
4 minutes
Grade Level: 8 - 12
Science Out Loud
Episode 8
5 minutes
Grade Level: 10 - 12
Science Out Loud
Episode 10
4 minutes
Grade Level: 10 - 12