The threat of Covid-19 left no choice for Japanese authorities but to postpone the 2020 Tokyo Olympics, and more sadly, hold the games with empty stands. Despite being sad news for the sports fans, this made the 2020 Olympics much more of an online event. It was now a much bigger field for IT companies to show off their advances. 

AI could be found everywhere in this year’s Olympics. From greeting athletes in the airport to helping the coaches analyze their team’s performance, there was always an intelligent agent at work. We are going to review some of them in this week’s post.

  • In the venues

From the instant the national teams land on Japanese soil, AI accompanies them on their trip. Haneda robotics lab robots greet athletes at the airport, check their baggage for illegal items, and show them the directions. There are chatbots in the venues providing helpful info: traffic, directions, schedules, scores,.. etc.

Since the Japanese know less English than the citizens of many other developed countries, several instant translation technologies are made available to the athletes and visitors. These include the app VoiceTera which translates Japanese to more than 27 other languages and vice versa, the wearable glasses made by NTT DOCOMO that use image processing and NLP to translate Japanese, and the palm-sized wearable translator by Panasonic.

Several self-driving vehicles designed by Toyota, an official partner of the 2020 Olympics committee are used during the games. Toyota’s “delivery support robot” moves up to 12 kg of baggage, also delivers food and drinks for athletes and fans (who maybe will be present in their seats in the next Olympics) using the mobile app. Toyota’s “human support robot” helps disabled persons with accessibility.

  • While training

3D athlete tracking technology (3DAT) made by Intel allows coaches to monitor every small detail of their trainee’s movements, using vision algorithms to specify the position of joints and muscles. They use the popular deep learning library Tensorflow to process the video input and detect body parts, clothing, and facial features of the athletes. 

Machine learning and image processing algorithms are a critical part of 3DAT. The ML gradually learns about each athlete and marks the important features and high-risk changes in his/her performance. The team behind 3DAT claims this AI system may be adapted to analyze the posture and movement changes to predict -and prevent- future soft tissue injuries. 

  • During the match

There is so much possibility for AI in sports events. An AI-based intelligent scoring system can be a helpful second-referee in gymnastics, and a robot that brings your tools and snacks is appreciated by every athlete, especially during a deadly pandemic.

It took Omega, the official Olympics timekeeper from 1932, four years to teach its AI beach volleyball. The timekeeper technology from Omega helps the fans with real-time details and technical data of the match. It combines live video from cameras with the information it receives from the sensors inserted in athletes’ clothing and processes them using machine vision and machine learning algorithms to produce real-time game information, including speed, direction, jumping height, pass, and technique type.

Toyota “field support robots” acts as an assistant staff in the matches. It can move objects from and into the field using the shortest route E.g. bring back javelins during the throwing events, bring the rugby ball to the center of the field before kick-off, … etc. This task was previously done using radio-controlled vehicles which needed skill and human work to do the job.