EP.40 A ROBOT CONTROLLED BY A FUNGUS

Amazon hires start-up founders, Google Deepmind impresses again, a coffee table as a robot & much more...

The fungus-powered robot 🦠

A team of scientists at Cornell University has developed a biohybrid robot that uses the nervous system of a fungus to drive its movements. The robot is equipped with an electrical interface that records the electrophysiological activity of the fungus's mycelium. This activity is then processed and converted into a digital signal that the robot can understand. The robot moves in response to the fungus, which sends out signals in response to environmental changes, like light.

The team built two versions of the biohybrid robot. One is a simple wheeled unit, while the other is a spider-shaped robot with soft legs. The spider robot is equipped with a Petri dish of fungus on top, which responds to light and other stimuli, sending signals to the legs to get moving. First, the robot moved based on natural continuous spikes in the signals from the mycelia. Second, the scientists shined ultraviolet light onto the fungus, which made the robot change its movement. Finally, the team demonstrated that they could override the fungus signal entirely if they needed to manually control the robot.

The potential applications of this technology are vast. The team envisions robots that can sense soil chemistry and decide when to add fertilizer, mitigating the downstream effects of agriculture. The fungus-powered robot could also be used in search and rescue missions, where it could navigate through rubble and detect chemicals.

The fungus-powered robot is a groundbreaking achievement in the field of biohybrid robotics. It demonstrates the potential of using living systems to improve the autonomy and responsiveness of robots.

---

A robot tying shoelaces 🥾

Deepmind is cooking again. From tying shoelaces to tightening screws, robots are getting closer to achieving dexterity that rivals our own. Two new AI systems, ALOHA Unleashed and DemoStart, are paving the way for robots to learn and master complex tasks.

ALOHA Unleashed is a new method that enables robots to learn complex tasks using two arms. This system builds on the ALOHA 2 platform, which was developed to improve dexterity in bi-arm manipulation. With ALOHA Unleashed, robots can learn to tie shoelaces, hang shirts, and even repair other robots.

DemoStart is a reinforcement learning algorithm that uses simulations to improve real-world performance on a multi-fingered robotic hand. This system requires 100 fewer simulated demonstrations to learn a task than traditional methods, making it a game-changer for dexterity research.

One of the biggest challenges in robotics is transferring knowledge from simulations to the real world. DemoStart's progressive learning approach automatically generates a curriculum that bridges this gap, making it easier to transfer knowledge from simulations to physical robots.

Isn't it exciting to see the progress being made in robot dexterity?

---

Coffee table that can walk 🪑

Imagine a coffee table that can walk around your living room, bringing you a drink or a snack. It sounds like science fiction, right? But for one maker, this was a challenge worth taking on. Meet the Carpentopod, a walking table resulting from years of experimentation and innovation.

The Carpentopod's unique leg linkage was designed using software that evolved thousands of virtual generations of leg variations. The goal was to create a linkage that could walk smoothly and efficiently, with a focus on minimizing bobbing and foot slips. The result is a design that's both functional and aesthetically pleasing.

The robot’s creator used Autodesk Fusion 360 to design and test the table's components, including the curved belly and legs. The table was then built using a combination of CNC machining, sanding, and lacquering. The result is a beautiful and intricate piece of furniture that's as functional as it is decorative. The robot is powered by two geared 24V brushless motors, which are controlled by an Arduino board and a Bluetooth module. The result is a table that can walk around the room with ease, bringing a new level of interactivity to the living room.

The Carpentopod is more than just a clever gadget - it's a masterclass in engineering and innovation. From the design of the leg linkage to the construction of the table itself, every detail has been carefully considered and executed. Whether you're an engineer, a maker, or just someone who appreciates clever design, the Carpentopod is sure to inspire.

---

---

Meme of the week 🤖

Alternative reality among robots.

---

---

Learn how to simulate mobile robotics virtually! 🦾

In this course, participants delve into the exciting realm of virtual robotics using MATLAB® and Simulink® Robotics Playground.

After completing this course, you will be able to:

• Setting Up and Programming: Establish a foundation by setting up MATLAB® and Simulink® Robotics Playground.

• Sensor Integration and Visualization: Program a variety of common robot sensors using Simulink®.

• Understanding Virtual Robotics: Acquire skills to operate within a virtual space.

---

Amazon hires Covariant founders ♻️

Amazon has hired three of the founders of Covariant, a Bay Area startup that develops AI for advanced warehouse robotics systems, in a deal that's being called a "reverse acquihire." As part of the arrangement, Amazon will receive a non-exclusive license to Covariant's AI models.

Covariant's technology automates warehouse tasks such as order picking and sortation, item induction, and depalletization. Amazon's move into warehouse robotics started with its acquisition of Kiva Systems about 11 years ago. Since then, the company has rolled out a series of warehouse robots across its operations, seeking to automate the process of moving products and packages through its fulfillment and sortation centers.

The deal with Covariant is the latest example of Amazon's efforts to expand its AI and robotics team in the Bay Area. With Covariant's AI models, Amazon aims to drive new ways to generalize how its robotic systems learn and provide dynamic opportunities for how it uses automation to make its operations safer and better deliver for customers.

The "reverse acquihire" model is becoming increasingly popular in the tech industry, with companies like Microsoft and Amazon hiring founders and employees from smaller startups and licensing their technology. This approach allows companies to bring in top talent and technology without having to acquire the entire company.

---