Gerardo Bledt

Companies large and small are seeking engineers with up-to-date, subject-specific knowledge in disciplines like computer engineering, automation, artificial intelligence, and circuit design. Mid-level engineers need to advance their skillsets to apply and integrate these technologies and be competitive. #renewables #power #energy #wind #offshore #onshore #solar #hydropower #floating #nuclear #geothermal #efficiency #transition #susteinability #circular #economy #esg #clean #decarbonization #electrification #climate #hydrogen #mobility #greensteel #ammonia #transportation #storage #fuelcell #smartgrid #der #powersystem #generation #transmission #distribution #electricity #machines #motors #drives #sensors #iot #education #engineering #innovation #entrepreneurship #creativity #technology #digitalization #cybersecurtity

5

What would be the best way to gather information about a field with both high throughput and high accuracy at the same time? My Ph.D. research addresses this question through our groundbreaking work in robot manipulation, computer vision, optimization, and aerial robotics. We transformed this moonshot idea into reality. Please check out the video below, presented by my advisor, Dr. Jin Jian, and our undergrad research assistant, Maximus Shurr. #robotics #precisionagriculture #phenotyping #agrobotics #computervision #drones #uav #digitalagriculture #agriculture

68

4 Comments

This is a snippet about a course I developed in the Robotics Dept at University of Michigan College of Engineering—ROB 311: How to build robots and make them move. What’s cool is the entire course and lectures are available on GitHub! https://lnkd.in/en4zt9kg

1,126

20 Comments

Hot off the PRESS! Our latest work on integrated finite elements with neural networks for solving thermoelasticity problems has been accepted in Computational Mechanics, and it is now available online: https://rdcu.be/dFPbt. Here we train and use a temporal convolution network to predict the temprature variations over time, and use the solution as a material point input within a classical elasticity solver; thus we are reducing the computational cost of the problem. This project is a part of our endeavours to introduce I-FENN to accelerate FEM using ML, while preserving the quality of non-linear and multiphysics FEM solvers. Kudos to @Diab Abueidda for leading this amazing project! Stay tuned for more interesting work, appearing SOON! #mechanics #ml #research #thermal #elasticity

84

7 Comments

✨ Diablo Saukele 🤖 💡 Direct Drive Technology Limited Diablo robotics has collaboration with CMU @Carnegie Mellon University Professor Zhang, Ji for the Diablo Saukele Navigation, SLAM! 🚀 Excellent compatibility: The software design is not heavily dependent on developing environment, and you can use it on most Linux devices that support serial ports. 💻 Simple control interface: We provide the robot visualization capabilities of rviz2 and gazebo in the function package of ROS2, which can be controlled by a customizable remote control plug-in. 🛡 Better safety: The system is equipped with a variety of safety assurance logics, which ensures safety to its best ability under various special conditions in the development process. 📽 Watch the video and witness Diablo's impressive SLAM technology #DirectDriveTech #Diablo #Robotics #CMU #SLAM #innovation 👉 Follow me Levent Coskun to stay up to date on #robotics #AI #hightech and #innovation. To get notifications about my posts please activate the notification bell 🔔 in my profile

51

26 Comments

The most common question that comes up during my 1-on-1s with aspiring roboticists is: "I've completed my undergraduate in XYZ field and now I want to build my career in robotics, where do I go from here?" This is when I need to backtrack a little. Robotics is an interdisciplinary field with numerous specializations. To help my mentees figure out the "where next" part of their career, I need insights into their past career journey and what their dream in robotics looks like. For instance, do they want to sketch out custom robots and then design and fabricate them? Or, just sketch and simulate novel robot designs and leave fabrication to others? But what if you want to tinker more with circuits and PCBs? Is AI and Deep Learning on your mind? Or, do you simply want to focus on algorithms for mobility, speech, vision, etc.? During the sessions, I give the mentee an opportunity to speak their heart out first. Where they currently are, how they got there, and where they want to get to. I help them with seed ideas but only enough not to bias them into one realm of robotics over the other. This is the tricky part of the session. Not everyone opts for career counseling early on. Sometimes a mentee may have pursued a slightly different field in undergraduate and is struggling to connect the dots to robotics. The thing with undergraduate studies being a broad exposure to the mother branches like Electrical, Mechanical, etc., is that you might think you need a 180-degree pivot, but it may not always be necessary. Need help figuring out your career trajectory into robotics? Feel free to connect and let's chalk out a plan that helps you achieve your dream career.

7

Exciting to see Carbon's lattice simulation development highlighted on this recent NVIDIA blog post!

21

1 Comment

Ready to explore the realms of signals and systems? Check out "Signals and Systems with MATLAB and Simulink," a textbook designed to offer a thorough yet accessible approach to problem-solving. 📚 Ideal for students poised to make their mark in engineering studies and careers. 🚀 #MATLAB #Simulink #SignalsAndSystems

9

#AdditiveManufacturing put to work at sea When a team of #engineers from the Consortium for Advanced Manufacturing Research and Education (#CAMRE) loaded their 3D hybrid-metal printer onboard #USSSomerset as part of the experimentation sector of Exercise Rim of the Pacific 2024, they had no idea that they would soon be asked to solve a real-world engineering casualty. Hours after being loaded on board, a critical component of the #reverseosmosis #pump, which generates #cleanwater for the crew – an absolute necessity for #ships spending long periods at sea – shattered. “What we didn’t expect was that we would have the opportunity to directly help ship readiness so soon,” said Lt. Charles Wallace in a Navy press release. Wallace is a mechanical engineer from the Naval Postgraduate School and was one of the team members onboard. “Especially for something as mission-essential as a reverse osmosis pump, where if you run out of water you’re going to be coming home pretty quick.” While 3D printing on Navy ships is still in its infancy, Somerset is not the first ship to utilize AM. In April, 2024, the amphibious transport dock #USSSanDiego (LPD 22) piloted a liquid metal jetting additive manufacturing process fielded by the CAMRE team, operationally showcasing this novel technology’s capabilities at sea. What makes the Somerset demonstration unique is that the machine is a metal hybrid design, combining subtractive and additive manufacturing in one machine. Before they tested the model on Somerset, this meant alternating between different machines, however by combining the two processes it effectively streamlines the overall workflow. Summer 2024 -- #USNavy #Technology #RIMPAC

1

1 Comment

From Part 2 of my Waymo Emails: Not many engineers pay attention to fluids and how they impact sensors, but Waymo does. In their blog post « Guided by the Flow » they show how using Computational Fluid Dynamics (CFD) simulations helps reducing drag and improving sensor longevity. It’s not a big deal on one trip with one car, but when preparing for mass production (as I believe Gen 6 of sensors is doing), it could mean a big difference over the years. #selfdrivingcars #robotics #lidar #computervision #deeplearning #ai

71

4 Comments

👉 Check out this new webpage containing MATLAB and Simulink Resources for students. This page includes helpful links to tutorials, documentation, examples, cheat sheets, and more! https://spr.ly/6040gLnLI

6

1 Comment

Why don't most robotics conferences use double-anonymous (aka double-blind) review? Many other fields do so successfully, what would be different in robotics? Two years ago my lab got interested in these questions. We spent a few months digging through the literature and discussing how it would work in robotics. The result of that effort was a literature survey/open letter that we circulated with the IEEE RAS leadership and others trying to encourage the robotics field to seriously consider changing our review process. I have finally gotten around to posting that document on arXiv so that everyone can read our summary and join the discussion: https://lnkd.in/eGv4bBtF The short version is that single-anonymous review has been shown repeatedly to be biased toward top PIs, universities, and countries. It is almost impossible to review a paper fairly when you know (or have heard of) the author/university vs when you have not. Interestingly, biases in terms of author gender, ethnicity, and other metrics are often not a major issue (but this may be field-specific). There was a great discussion of these issues at the ICRA Workshop "Debates on the Future of Robotics Research" this year. You can watch the video here: https://lnkd.in/eB98_NZR I want to thank Katherine J. Kuchenbecker for running a test at the IEEE-RAS sponsored World Haptics Conference last year, showing that it can be done in robotics. (Also, thank you for reminding me to post this on arXiv!) Also, thank you to Nihar Shah, also from CMU, for his advice and extensive work in this area: https://lnkd.in/eqWQNg9z Finally, thank you to the leadership at IEEE Robotics and Automation Society for engaging in discussions with my group on this. There is more to be done but I appreciate the openness to the idea that everyone showed. Now it is time to take the next step and try this in more of our conferences.

451

15 Comments

🚀 No more weeks spent writing CFD codes! 🔍 Our previous research shows that by creating the right programming abstractions, the process of writing CFD codes can be significantly accelerated while ensuring scalability on parallel computers. Check out our open access results for all the details! 📖://https://lnkd.in/ez-EWrpU

3

I've shared my view about GPUs for #CFD. It is evident that speed-up is definitely amazing, and even an inefficient algorithm is sometimes 5X faster than the CPU. Of course, as bigger GPUs hit the market, we can run industrial-oriented CFD workflows in GPUs faster; that is great. But, I still wonder if problems commonly encountered in industrial applications can be solved with an affordable GPU solution. I don't know. Some #CFD algorithms are meant for something other than GPUs, even in cartesian frameworks. We know that, but X86 systems are cheaper than GPUs, and we have lived with it. We know that our CFD kernels are, most of the time, fetching stuff instead of crunching numbers. If you think about unstructured #CFD, things can rapidly go south. Anyway, I recently read about NVIDIA Grace, which is an ARM CPU, which I found interesting due to the great characteristics of ARM. I don't know; I'm just sharing my thoughts here. Probably, I'm wrong, which is usually the case with #HPC. https://lnkd.in/e2uReFsP Happy Computations and "Muchas Gracias"

71

13 Comments

I generated a synthetic EEG signal with Matlab and performed FFT analysis.

3

1 Comment

A few weeks ago I wrote a guest post for the WorQuick blog on incorporating Python, Jupyter Notebook, and SymPy Mechanics into an engineering workflow. As I put that post and the accompanying tutorials together, I found myself creating several graphics that I found myself to be very proud of, showcasing some of the 2D and 3D modeling I've been practicing lately, Python plotting, and animations with Mobile Multibody Dynamics. Here is a quick summary of some of that eye candy. #python #mechanicalengineering #dynamics #jupyternotebook #blender https://lnkd.in/eRt2c8F6

3

Impressive video generated by Luma with the prompt "Generate a corner case in autonomous driving: a car cuts in." Utilizing video generation technology for autonomous driving testing is highly promising!

10