Posts

SE2 & Forward Kinematics

In our previous blog on robotic grasping , we briefly broached the topic of Inverse Kinematics (IK) and Jacobians. In this blog, we will establish some fundamentals about the mathematics behind Robot Arm Kinematics that will allow us to dive deeper. The SE(2) Group  The Special Euclidean group (SE2) [group being a mathematical set and operations that satisfy some laws] can be used to represent the rigid transformations of a planar object, a 2D arm, on a plane. We are going to first use a 2D arm, composed of a set of joints and links and see how we can make its end-effector reach certain parts of the 2D workspace, before going on to see if we can do the same in 3D (the space occupied by real-world robotic arms). Concretely, each joint of the 2D arm has a 2D coordinate frame attached to it, which is described relative to the 2D coordinate frame of the previous joint, all the way down from the end effector to the base of the robot. These coordinate transforms in 2D space can be repres...
Post a Comment
Read more

Unlocking Robotic Grasping: The Fusion of Perception, Mathematics and Control

Image
 Imagine how you reach out to an object within your field of view, extend your arm to it, and grasp it within your fingers. This task, effortless for us, is rather more complicated for robots. Let's dive into the fascinating world of robotic grasping, where mathematics meets mechanics to mimic human manipulation. 6 DoF pose estimation using NVIDIA's  Deep Object Pose Estimation  neural net The Perception Challenge : At its core, robotic grasping begins with perception. Just as our eyes and brain work together to identify objects and their positions, robots use advanced computer vision algorithms to detect the 3D pose of objects. These algorithms act as the robot's "eyes," providing crucial information about what to grasp and where. Training robots to grasp objects involves using a perception model or computer vision algorithm to detect the 6 DoF (Degrees of Freedom) 3D pose of the object. This is often achieved using sophisticated neural networks like NVIDIA's...
Post a Comment
Read more

Teaching Robots to Do the Dishes: A Data-Driven Approach

Image
 Imagine a future where robots handle everyday chores like dishwashing. No more dishes piling up in the sink. No more having to load and unload the dishwasher. Well, a professor from New Jersey, Usman Roshan and his CTO at  7XR , have been working on just this, for the past few months.  Autonomous robot dishwasher (ARD1) identifies plates, tap, water, sponge, sponge box, dirty vs clean plate, which plate to pick, as we see in the four views below - all of this helps it to understand what it is doing  #robots   #robotics   #AI   #deeplearning   #machinelearning …  pic.twitter.com/whSFr8Bh03 — Usman Roshan (@Deeplearner2)  September 26, 2024 The X-post from  7xr.tech  showcases a glimpse of our robotic future, featuring a couple of robotic arms from  Elephant Robotics  tackling dirty dishes. The system is completely vision based, with cameras near the end-effectors and learns from 60 videos of the CTO washing dishes at ...
Post a Comment
Read more

A pocketful of Sun in Livermore

Image
The Sun and other stars are powered by nuclear fusion. 2 atoms of hydrogen come to a violent union to produce a helium atom, a neutron and some energy. The temperatures and pressures in the centre of the Sun were produced, if only for a few nano-seconds at the National Ignition Facility (NIF) at Lawrence Livermore National Labs (not too far from my home) on the 5th of December, 2022. It had been about 70 years in the making, and the net energy gain achieved was hailed by some as a breakthrough as revolutionary as landing on the Moon.  Solar temperatures and pressures are required to overcome the Coulombic or electro-static repulsion of the hydrogen atoms (really hydrogen isotopes deuterium and tritium - and their fusion is called D-T-fusion) and get their strong nuclear forces to take over and bind the constituent D-T protons and neutrons into one resulting Helium atom.  There are several ways to achieve this temperature and pressure. The first one, is inside a thermo-nuclear ...
Post a Comment
Read more

Christmas Toy

Image
Over the Christmas holidays, I decided to use my nascent electronics skills to rustle up a little toy for the Christmas tree. I took inspiration from the little Asian waving cat, Maneki-neko , that's often seen at Asian stores and restaurants. I wanted to make a little Santa figurine that would light-up, wave and say "Merry Christmas" when someone came close to it. I decided to use the Arduino Uno (because I already had one) for the microcontroller (uC), an  addressable light strip for the lights, an HC-SRO4 ultrasonic sensor for proximity-based triggering and a tiny 4-ohm 3-watt speaker  for the sound to come out of.  The first thing was to make the setup say something. I found out, initially from this youtube channel , that there is an Arduino PCM library that can play short audio samples. These samples are encoded in the Arduino sketch as an array of numbers. HiLowTech describes the process of converting an audio sample into this series of numbers. Downsampling the...
Post a Comment
Read more

Inventing Cheap Catalysts by Modeling Atomic Interactions using Graph Neural Nets. How ML can help save the world from catastrophe.

Image
Computational-chemist - the new breed of scientist In today's rapidly evolving tech landscape, a new breed of scientist/engineer is emerging: the computational chemist, along with their counterpart in biology and physics. Their mission? To harness the power of Machine Learning (ML) to expedite the discovery of revolutionary new compounds. These compounds may serve as catalysts in planetary-scale CO2 removal machines, or as key components in ground-breaking battery chemistry that might someday allow commercial airliners to go electric! Alternatively, they could form the basis for innovative drugs or vaccines in the fight against plague and disease. In this post, I discuss how Graph Neural Nets (GNNs) can be used to accelerate the discovery of new materials critical for fighting climate change. At the recent CVPR conference , the premier gathering for scientists and engineers in Computer Vision and Machine Learning, we were told by Larry Zitnick , the co-architect of the celebrated M...
Post a Comment
Location: San Francisco, CA, USA
Read more