Seven robots you need to know

Pointing the way to an android future

Walking. Grasping an object. Empathising. Some of the hardest problems in robotics involve trying to replicate things that humans do easily. The goal? Creating a general purpose robot (think C-3PO from Star Wars) rather than specialised industrial machines. Here are seven existing robots that point the way towards the humanoid robots of the future.

Photo: Getty

Atlas

Use: Originally built for Darpa Robotics Challenge
Made by: Boston Dynamics
What it tries to do: Achieve human-like balance and locomotion using deep learning, a form of artificial intelligence.

“Our long-term goal is to make robots that have mobility, dexterity, perception and intelligence comparable to humans and animals, or perhaps exceeding them; this robot is a step along the way.”​

Marc Raibert, founder, Boston Dynamics

Features:
• 1.7m tall and weighs 82kg
• Can walk on two feet and get back up if it falls down
Human equivalent: Legs/skeleton/musculature

Superflex

Use: Military. Part of Darpa’s Warrior Web project
Made by: SRI Robotics
What it tries to do: A suit that makes the wearer stronger and helps prevent injury

Superflex is a type of ‘soft’ robot, which can mould itself to the environment or a human body in a way that typical robots can’t. The goal is to make machines that feel and behave more like biological than mechanical systems, and give additional powers to the wearer.

Features:
• Battery-powered compressive suit weighs seven pounds
• Faux ‘muscles’ can withstand 250lb of force
Human equivalent: Musculature

Photo: SRI International

Amazon Echo

Use: Voice-controlled speaker
Made by: Amazon
What it tries to do: Lets you control devices by talking to them

It may not have any moving parts, but Amazon’s Echo – and Alexa, the digital assistant that lives inside it, is definitely trying to solve one of the central problems in robotics: how to create robots that can recognise human speech and provide natural voice responses.

You can tell Alexa to:
• Control your light switches• Give you the latest sports scores
• Help tune your guitar
Human equivalent: Voice and ears

Life-like humanoids

Use: Natural interactions
Made by: Hiroshi Ishiguro Laboratories
What they try to do: Create a sense of ‘presence’, or sonzai-kan in Japanese, by making robots that look identical to humans

“Our goal is to realise an advanced robot close to humankind and, at the same time, the quest for the basis of human nature.”

Hiroshi Ishiguro

Human equivalent: Feelings and emotions

First image shows ‘Geminoid-F’, video shows ‘Erica’ (Erato Ishiguro Symbiotic Human-Robot Interaction Project)
Geminoid-F photo: Getty, video: Hiroshi Ishiguro Laboratories.

Pepper

Use: Day-to-day companion, and customer assistant
Made by: SoftBank
What it tries to do: Recognise and respond to human emotions

While Pepper clearly looks like a robot rather than a human, it uses its body movement and tone of voice to communicate in a way designed to feel natural and intuitive.

See Pepper's visit to the FT

Human equivalent: Feelings and emotions

Photo: Getty

Robo Brain

Use: Knowledge base for robots
Made by: Cornell University
What it tries to do: Accumulate all robotics-related information into an interconnected knowledge base similar to the memory and knowledge you hold in your brain.

The human brain is such a complex organ that it would be extremely difficult to create an artificial replica that sits inside a robot. But what if robots’ ‘brains’ could exist, disembodied in the cloud? Robo Brain hopes to achieve just that.
Researchers hope to integrate 100,000 data sources into the database.

Challenges: Understanding and juggling different types of data

Google Car

Use: Self-driving car
Made by: Google
What it tries to do: Group learning and real-time co-ordination

The true ambition behind Google’s automotive efforts is not just to make a car that can drive itself. Instead, it’s to use group learning to strengthen artificial intelligence, so that if one Google car makes a mistake and has an accident, all Google cars will learn from it. This involves managing large-scale, real-time co-ordination.

What happens when robots rule the road

Photos: FT Graphic/Getty/Dreamstime