Why is it important: Since Boston Dynamics first introduced the famous legged robot in 2013, many organizations have introduced machines that autonomously carry out a variety of professional and recreational activities. The latest research uses badminton to develop robot agility and coordination. This will help you improve your training method and ease of use.
Scientists at EthZürich recently published a study and video (below) explaining how they trained Quadrupedal robots to play badminton. This study may demonstrate the usefulness of a full-body training model that does not divide activities into individual steps, and may expand the scope of learnable tasks.
The standard ANYMAL-D robot was equipped with Dynaarm and taught using standard reinforcement learning models with minimal software or hardware specialization. Additionally, the machine has learned to track Shuttlecock using only onboard stereo cameras.
https://www.youtube.com/watch?v=2dk8dsrlhli
The robot has learned to navigate courts using NVIDIA's ISAAC gym virtual simulator. This took about 7,500 iterations in almost 5 hours using the RTX 2080 TI graphics card. Furthermore, the noise prediction model helped track Shutttlecock when it escaped from the limited field of view of the camera.
Eventually, I learned to develop a variety of complex movements, including starting motion, backswing, hits, running, recovery, and more, without the need for researchers to teach them individually. He adjusted his movements based on the shuttlecock's perceived speed, galloping across the court and bumping into him in time.
This experiment demonstrates the adaptability of Anymal-D, designed to perform inspection, emergency response, and other tasks in a variety of industrial environments. The company's website shows robots navigate indoor facilities, with indoor equipment, gravel, stairs, caves, snow, mud and flooding operating 24 hours a day. The upcoming model, Anymal-X, is certified for a more dangerous environment to help carry out testing that includes oil, gas and chemicals.
Badminton is the latest leisure activity used to demonstrate the adaptability of foot robots. Boston Dynamics famously showed the dance and gymnastics performances of the original Atlas model, and the company's four division spot units showed how robots could one day help stage productions.
Meanwhile, Google Deepmind trains robot arms to play table tennis, Engineered Arts introduces humanoid robot artists, and the NBA recently deployed robots to help athletes practice. Chinese companies recently deployed robots to compete in half marathons and kickboxing tournaments.
