Advertisements
In action: images of the bike show that he is navigating a racetrack without a cyclist at the helm - instead, he follows commands and responds to stimuli

If you have trouble cycling, scientists may have the answer! AI bike developed in China self-balances, avoids obstacles and responds to voice commands

  • Tsinghua University researchers used a small AI chip mounted on the rear wheel
  • Microphone captures voice commands and speed sensor detects momentum
  • Breakthrough comes three years after an April Fool’s video from Google
Advertisements

We already have cars without a driver, but scientists have now made a self-cycling bicycle.

Advertisements

Perfect for the wobbly cyclist, the bike keeps itself upright, sees speed bumps and automatically winds around obstacles.

The bike is powered by artificial intelligence, so like voice-controlled assistant Alexa, it responds to commands to speed up, slow down and turn left or right.

It was made by researchers at Tsinghua University in Beijing, using a small AI chip on the rear wheel.

The breakthrough comes three years after an April Fool's spoof video from Google suggesting that the company had made a self-driving bicycle, but a "making of" video showed it was fake.

In action: images of the bike show that he is navigating a racetrack without a cyclist at the helm - instead, he follows commands and responds to stimuli

Advertisements

In action: images of the bike show that he is navigating a racetrack without a cyclist at the helm – instead, he follows commands and responds to stimuli

AI: The bike is powered by artificial intelligence, so like voice-controlled assistant Alexa, it responds to commands to speed up, slow down and turn left or right

AI: The bike is powered by artificial intelligence, so like voice-controlled assistant Alexa, it responds to commands to speed up, slow down and turn left or right

AI: The bike is powered by artificial intelligence, so like voice-controlled assistant Alexa, it responds to commands to speed up, slow down and turn left or right

HOW DOES IT WORK?

The bicycle was made by researchers at Tsinghua University using a small cross-paradigm neuromorphic chip called Tianjic, which mounted on the rear wheel.

A gyroscope is used to measure the stability of the bike to keep it upright, while a camera detects obstacles in its path.

Advertisements

A microphone then captures voice commands from the rider and a speed sensor detects how fast it goes.

Experts say that the technology behind the bike, which also has the ability to perform S-turns and high-speed turns, works in the same way as the human brain to find and respond to obstacles.

But it also uses machine-learning algorithms, which means that every cycle function can be operated from a single chip.

The Chinese researchers use a gyroscope to measure the stability of the bicycle to keep it upright and a camera to detect obstacles in its path.

A microphone picks up voice commands from the rider and a speed sensor detects how fast it goes.

Advertisements

Professor Will Stewart, Chairman of the Communications Policy Panel of the Institution of Engineering and Technology, who was not involved in the research, said: & # 39; This is a fascinating example of a multifunctional programmable AI chip, which cleverly demonstrates its many functions needed to drive a bike.

"It is controlling the balance, understanding what stands in front of the bike and also steering.

"It is a step forward to make AI much more useful for our daily lives."

A study of the bicycle, published in the journal Nature, shows that it is perfect for cross-country bike rides, being able to follow a person in front and follow it in a straight line. & # 39;

Avoiding Obstacle: Experts say that the technology behind the bike, which also has the ability to perform S turns and fast turns, works in the same way as the human brain in finding obstacles

Avoiding Obstacle: Experts say that the technology behind the bike, which also has the ability to perform S turns and fast turns, works in the same way as the human brain in finding obstacles

Advertisements

Avoiding Obstacle: Experts say that the technology behind the bike, which also has the ability to perform S turns and fast turns, works in the same way as the human brain in finding obstacles

Study: analysis by bike, published in the journal Nature, shows that it is perfect for cross-country bike rides, being able to follow a person ahead and follow them in a straight line

Study: analysis by bike, published in the journal Nature, shows that it is perfect for cross-country bike rides, being able to follow a person ahead and follow them in a straight line

Study: analysis by bike, published in the journal Nature, shows that it is perfect for cross-country bike rides, being able to follow a person ahead and follow them in a straight line

In a video on YouTube you can see how you cycle over a speed bump and avoid traffic cones, without a rider operating the pedals.

A researcher walking next to the bike simply instructs him to accelerate, to go to the left or to cycle in a straight line, which he does with the help of his audio sensor and motor.

Advertisements

Nick Chamberlin, policy manager at British Cycling, said: "We are very curious about the benefits of AI technology to enable more people to enjoy the benefits of cycling, especially the elderly and those with a life-limiting illness or disability.

"New technology means that the future of mobility is exciting for the first time in a generation.

"There are still many questions about the safety of all autonomous vehicles, but as we learn more, we can understand the positive benefits and work with society and policy makers to ensure that those benefits are realized."

Experts say that the technology behind the bike, which also has the ability to perform high speed S turns and turns, works in the same way as the human brain to find and respond to obstacles.

But it also uses machine-learning algorithms, which means that every function of a bicycle can be operated from a single chip.

Advertisements

Details of the research were published in the journal Nature.

. [TagsToTranslate] Dailymail