Teaching robots to handle various door types is easier said than done. Robots that are not trained properly can fail to grasp and open doors they have not encountered before. This video from IHMC Robotics shows Navia traverse many types of doors autonomously. The robot was tested with pull and push doors with 4 types of opening mechanisms
[HT]
This is the KAERI Armstrong: a hydraulic robot with 100kg single arm payload capacity. It is capable of handling a large amount of reaction force and torque generated when operating an impact wrench. As this video shows, it can perform fasten bolts effectively:
Torobo is a humanoid robot from Tokyo Robotics for machine learning research and task automation. The new version is lighter than the older version and has faster motion. It is expected to become available in March of 25. This torque-controlled humanoid robot has 2 x 7DOF arms and a 3DOF torso.
Here is another mobile platform for dealing with rogue drones. The Cerberus XL C-UAS has a rugged construction and can be deployed quickly to deal with threats. It has radar, EO/IR, and RF detection. It works with a variety of sensors. It can detect and track over 500 targets simultaneously.
The Cerberus XL C-UAS has a modular design. It can be configured for air, ground and maritime threats. This can be put to use in extreme conditions. The Teledyne Cameleon C2 interface allows for camera control, mission recording and mapping.
[HT]
In the past few years, we have covered a few robots that can juggle a few balls. It is not easy to program robot arms with limited speeds to pull it off. This video from Daniel Simu takes a different approach. It shows how humans and robots can collaborate when it comes to ball juggling.
As Daniel explains, he 3D printed a hand to pull this off. You can find out more here.
Keyframing is a technique used in animation to start the start and end points of an action and have the computer fill the in-between frames. Disney Researchers have now come up with RobotKeyframing, which is a new way to control legged robots. To guide the robot’s learning process, the framework uses a multi-critic reinforcement learning algorithm for balanced handling both dense and sparse rewards.
This framework enables robots to anticipate future goals and improves their ability to reach their goals.
[HT]
In the future, we will have vehicles with flying modules that give you the best of both worlds. XPENG AEROHT’s Modular Flying Car is based on this approach. It has a ground module that carries the air module. These two automatically separate and combine. The ground vehicle has a 6×6 all-wheel drive system and is ready for off-road action.
The “Land Aircraft Carrier” air module completes its first manned flight!#XPENGAEROHT #FlyingCar #Tech #Innovation pic.twitter.com/cBkUIhd846
— XPENG AEROHT (@XPENG_AEROHT) October 29, 2024
The human piloted air module is an electric vertical and takeoff landing system. It has manual and automatic dual driving modes. It also features a 270-degre panoramic 2-seater cockpit.
[Ht]
Here is one of the coolest humanoid robots money can buy. The Booster T1 stands 1.2m tall, weighing 30kg. It can do push ups, walk, perform Kung Fu and play soccer. Thanks to its fall recovery system, it can recover. The T1 has a circular 6-microphone array and speaker. It has a depth camera and LiDAR.
Robots are going to play big roles in our lives in the future. They will need to be able to learn new skills all the time. With AI, they will be able to do much faster. This video from Unitree shows how robots can learn human-like natural walking in 2 days.
In order to get better at tennis, you need to work on your forehand, backhand and other shots all the time. The PongBot Pace S can help with that. It is an AI powered tennis robot with a position sensor that can adapt to your level of play. It comes with 300+ drills and has room for 15 balls. You get a visual interface on your mobile phone for custom drills.
