EP.33 ROBOT WALKING ON THE GREAT WALL
Amazon's delivery drones and AI, an extra robo-thumb & much more ...
European open-source robot gets smarter! 🧠
Hugging Face, an open-source AI code community and Pollen Robotics have collaborated on their first project: an open-source robot called Reachy2 that can perform household chores. Reachy2 is a humanoid robot designed by Pollen Robotics, and Hugging Face trained it to interact safely with humans and dogs using a novel machine-learning algorithm.
The training process involved a human wearing a virtual reality headset to teleoperate the robot, which was then studied by the algorithm to learn how to perform tasks such as placing cups on a dish rack and handing objects to people. The dataset and model used for the demo are now open-sourced and available for anyone to use.
This development marks a significant step in the advancement of open-source robotics, with Hugging Face's Le Robot program offering technology essentially for free. The collaboration between Hugging Face and Pollen Robotics is a natural alliance, with both companies sharing a commitment to open-source and accessible technologies.
Pollen Robotics has a rich history of developing open-source technologies for real-world applications, and its Reachy robots are designed for research and interaction with humans. The upcoming Reachy2 robot promises to be a significant improvement, with all-new 7-DoF bio-inspired arms and a 3 kg payload capacity.
Amazon’s delivery drones received approval! 🥏
Amazon Prime Air has received approval from the Federal Aviation Administration (FAA) to extend its beyond-visual line-of-sight (BVLOS) drone delivery operations. This milestone will enable the company to expand its package delivery services, leveraging its onboard detect-and-avoid system to ensure safe navigation.
To achieve this approval, Amazon developed and refined its detect-and-avoid system over several years, submitting detailed engineering information to the FAA and conducting flight demonstrations to showcase safe navigation in real-world scenarios. The company's vision is to create a safe and scalable way to deliver packages to customers in 30 minutes or less using highly autonomous drones.
With this approval, Amazon Prime Air will expand its drone delivery area in College Station, Texas, and plans to integrate drone deliveries into its delivery network later this year. The company remains committed to safety, ensuring compliance with strict safety standards as it expands its delivery reach. This development marks a significant step towards Amazon's goal of delivering 500 million packages per year by drone by the end of the decade.
The first robot to climb the Great Wall! 🇨🇳
Robot Era, a Chinese company, has released a video showcasing its humanoid robot, XBot-L, climbing the Great Wall of China. This achievement marks a significant milestone, as XBot-L is claimed to be the first humanoid robot to accomplish this feat.
The video demonstrates XBot-L's ability to navigate the wall's uneven surfaces, broken pavement, and stairs with ease. It also showcases the robot's ability to perform tai-chi moves and adapt to dimly lit guard towers. According to Robot Era, XBot-L's proprietary reinforcement learning (RL) technology enables it to recognize complex road conditions and adjust its walking stance in real time.
While the video highlights XBot-L's advanced navigational and balancing systems, it also reveals a humorous moment where duct tape is visible around the robot's midsection. The purpose of the duct tape is unclear, but it adds a touch of humanity to the otherwise impressive display of robotic capabilities.
This promotional effort by Robot Era serves as a clever way to normalize the presence of humanoids in everyday life. As humanoid robots become more prevalent, it raises interesting questions about their integration into society, such as whether they will be treated differently at tourist attractions.
Meme of the week 🤖
Hey, listen robot. Take my job. Never my leisure.
Build robots with Applied AI! 🧬
After watching this live stream, you will be able to:
Design, test, and deploy AI applications for robots, Cobots, AMRs, UAVs, etc.
Streamline development with Model-Based Design tools.
Insights Include:
Command robots with AI chatbots (ChatGPT, Bard).
Detect and classify objects using deep learning.
Efficient motion planning and control with reinforcement learning.
Deploy CUDA-optimized ROS nodes for deep learning.
Getting an extra thumb! 👍
Researchers from the University of Cambridge have developed a wearable device called the Third Thumb, which is a controllable, prosthetic extra thumb that can be worn on the opposite side of the palm to the biological thumb. The device is controlled by pressure sensors placed under each big toe or foot and can be used to pick up and manipulate objects.
In a recent study, the team tested the Third Thumb on a diverse range of participants, including people of all ages and backgrounds, at the Royal Society Summer Science Exhibition. The results showed that almost everyone was able to use the device straight away, with 98% of participants successfully manipulating objects using the Third Thumb during the first minute of use.
The Third Thumb has the potential to benefit people with disabilities, as well as healthy individuals who want to enhance their productivity. The device could be used to perform tasks that are challenging or impossible to complete with one hand, or to perform complex multi-handed tasks without having to coordinate with other people.
Amazon hired an AI for damage detection! 🔎
Amazon has introduced Project PI, an artificial intelligence (AI) system designed to detect damaged goods and other issues before shipping them to customers. The system uses computer vision to inspect packages and flags any items that appear damaged, which are then reviewed by Amazon associates. This technology aims to improve the customer experience, reduce unwanted returns, and minimize waste and carbon emissions.
In addition to on-site AI tracking, Amazon is also deploying a Multi-Model LLM system to analyze customer feedback and returns, identifying potential issues in its warehouses, such as mislabeled products or bad batches. This system will help small and medium businesses on Amazon's platform and provide an extra layer of quality control for associates.
While Project PI is a positive development, it is unlikely to address issues specific to PC component buyers, such as receiving counterfeit or damaged products. The system only scans exterior packaging and does not inspect the actual product, so it would not detect issues like a putty-filled GPU or fake graphics cards.
Keep reading with a 7-day free trial
Subscribe to we all are robots to keep reading this post and get 7 days of free access to the full post archives.