Anselm Joey Mak

Teach robots how to autonomously navigate their environments with MATLAB and Simulink

347

16 Comments

I put together this brief comparison loosely and tightly coupled INS/GPS integration techniques. A quick guide for anyone working in autonomous systems or navigation in general.

63

🎇 On National Day, I went for a leisurely drive in San Francisco and ended up "stress-testing" a Waymo self-driving car on the road. 🚗 As an autonomous driving practitioner, who wouldn't be curious about the real-time robustness of the cutting-edge Waymo One driving system? While cruising downtown, I accidentally noticed a Waymo car tailgating me. While this isn't unusual for SF citizens, a wild, "evil" idea suddenly hit me: Why not directly "adversarially attack" the world's autonomous driving status quo? I executed an unexpected maneuver—suddenly reversing—to see how it would react. 🌟 The response was stellar! At the moment I reversed, Waymo One honked instantly—quicker than any human could—activated its hazard lights, and backed away to maintain a safe distance. This reckless move on my part served as an edge case to test the algorithm's robustness under extreme conditions and, potentially, could be a challenging training sample to enhance Waymo's future autonomous systems. 💫 Thrilled to personally experience how current cutting-edge autonomous algorithms handle rare driving behaviors—and how stable and safe Level 4 autonomy is in dealing with diverse scenarios. However, it also prompted deep reflection as an AI researcher in this field: 🤔 In an industry with little room for error, how can we ultimately avoid or minimize issues that AI fails to handle? 💡 I believe two research directions are particularly promising for achieving Level 5 level autonomy in future mobility systems: - 1️⃣ Development and deployment of vehicle-to-everything (V2X) cooperative systems (including V2V, V2I, V2P, etc). Our initial studies (e.g., V2X-ViT, ECCV'2022 arxiv.org/abs/2203.10638) show that in scenarios with severe occlusions or noise, such cooperative systems can significantly enhance the robustness of perception systems, thereby eventually improving traffic safety. - 2️⃣ Adversarial scenarios generation (including Sim-to-Real, generative modeling). Research done by my colleagues at UCLA (V2XP-ASG, ICRA'2023 https://lnkd.in/gu5nKVHD) shows that adversarial learning techniques can effectively simulate adversarial scenarios, greatly improving model robustness in complex "corner case" situations. Of course, it's often infeasible to collect such collision scenarios. 👨‍🏫 As a new Assistant Professor of CS at Texas A&M University, I will lead a group focusing on these exciting research directions, which can be a proactive approach to reducing accidents and improving safety for all. 🔥 I look forward to future collaborations with governments, academics, and companies to research and develop data and algorithms that can help enhance the safety of vulnerable road users, especially seniors and children. We envision a people-centered intelligent transportation system in the future. Interested in these topics? Let's connect and discuss further! #AutonomousVehicles #AI #MachineLearning #SmartCities #Transporation #Humanity #Mobility

144

3 Comments

I'm thrilled to announce the open-source release of UnitcellHub (https://lnkd.in/gBM9SuHH), a powerful lattice design and modeling suite I developed while working on NASA's Mars Sample Return Mission.   What is UnitcellHub? Awarded JPL's 2024 Software of the Year, UnitcellHub is a comprehensive structural lattice design suite, enabling robust geometry creation, automated finite element simulation, property database management, and an intuitive design space exploration Graphical User Interface (GUI). The software is composed of three distinct but interconnected submodules: UnitcellEngine (https://lnkd.in/gFJQ-DY9) leverages state-of-the-art technology like implicit geometry representation, meshless finite element analysis, and homogenization theory to automate a robust pipeline from geometry definition to performance prediction. UnitcellDB (https://lnkd.in/gB2-Wg-6) is a database of performance metrics for more than 15,000 lattice point designs (generated by UnitcellEngine), spanning more than thirty types of unitcells and their many possible geometric realizations. UnitcellApp (https://lnkd.in/gREmTarP) is a desktop and web-based graphical user interface that brings together the geometry visualization capabilities of UnitcellEngine, data in UnitcellDB, and machine learning, to enable a user to select the ideal lattice geometry for their given application. Together, UnitcellHub provides an end-to-end framework to understand and design structural lattices.   Ready to explore? Give UnitcellApp a try at https://lnkd.in/gHh4exze to start designing today or install UnitcellEngine into your Python environment with "pip install unitcellengine" to being modeling and simulating custom lattices.   What's coming? In the coming months, I'll release mini-webinars on using the existing features and customizing the tools for your specific needs. Stay tuned for me announcements or reach out directly to be added to the mailing list.   Why Open Source? Much of the work that we engineers and researchers conduct utilizes amazing software that we access free of charge through opensource licensing. It is hard to image doing our work without access to tools like Linux, Python, NumPy, PyTorch, Jupyter, and so many more. This is my small way of giving back to the community.   #opensource #unitcellhub #lattice nTop Metafold 3D

301

23 Comments

👩🏽‍💻 Preparing for an interview at MathWorks? 📝 Here's what to expect! #interview #interviewtips #softwareengineering #engineeringthefuture #matlab #simulink

7

For those interested in trying out systems engineering, I found the MBSE Fundamentals course from Caltech Center for Technology and Management Education (Caltech CTME) to be an accessible and effective introduction. MBSE is a big field, and Cameo is a complex tool. This course gives instruction on both that's comprehensive enough to enable real-world work but not so deep that it's overwhelming.

27

5 Comments

Opportunity

3

Here is a demonstration of a C++ simulation I made while sick this thanksgiving break, it is a very basic implementation of collision and 3D space. Will work on expanding the physics aspect of it to better demonstration fluid dynamics in 3D space. The code can be found on my github, aneangel/3dParticleSim

21

2 Comments

Charles Lafage let's have a pokemon battle to resolve this. Contenders: * Imants Smidchens (my younger brother, sophomore student in mechanical engineering). He is given **2 hours** to design a solution using traditional CAD. * Charles Lafage. You are given **200 hours** and unlimited access to generative AI technologies of your choice to generate a solution. Outputs needed: * Full BOM and cost estimate * GCODE for all 3D printed parts (Prusa Mini, PLA, 0.4 mm nozzle, firmware 6.0.2) * All parts must be standard COTS hardware (screws, nuts, etc.) or 3D printed. * Assembly instructions with visuals (e.g. a technical drawing) * Documentation of the design process (screen recording or write up) Challenge rules: * Thinking is allowed before the timer starts. Timer starts when interacting with software. Timer does not pause. * Imants Smidchens is not allowed to use any generative/machine learning/neural network/etc. AI. * Charles Lafage is not allowed to manually operate any traditional CAD software (but is allowed to execute scripts generated by AI, including ones that call APIs in CAD software). * The design challenge/problem will be the top reply to this post, evaluated by number of "Like" reactions. * Voting for the challenge concludes on June 14th. * Submissions responding to the challenge are due by June 28th, as LinkedIn posts quoting/tagging this post. Evaluation criteria: * The parts and assembly instructions will be procured, printed, and mailed to the person who posed the design problem for assembly and feedback. (I'll front the cost, but it should be reasonable). * I will also personally build a copy of each design, documenting the process in depth. * A poll will be posted on LinkedIn asking engineers to evaluate which option better solved the design challenge, including: * Quality of full design data produced (BOM, cost, gcode, assembly instructions) * Quality physical prototypes (functionality, ease of assembly, robustness) * Meeting requirements (for outputs & engineering challenge) Imants Smidchens I choose you. Charles Lafage are you up for the challenge?

51

8 Comments

For the same reason implicit modeling excels at lattices, it can create complex arrays and patterns at virtually no extra cost. We're talking arrays of tens, hundreds, thousands, or more (and not just identical ones). The output can be combined with other geometry and modified as any other object would be. Design parameters (not just offsets!) can even be varied throughout space within the array while maintaining the same computation speed. #implicitmodeling is changing what geometric complexity in CAD means and I can't wait to help push the frontier further. I decided to start making this simplified scale model of an adjustable exhaust nozzle as an experiment, in nTop, from scratch. It's inspired by the F-15 design and I've had a super fun time building it out. The current iteration is fully parametric, meshes super fast, and even includes a basic encoding of the main linkage with two angle inputs. This way, we can run (meshless!) CFD on it in different configurations. If I want to export it to other software, I have several options, and the .implicit file is just 123kB. What new #implicitmodeling features for arrays and patterns would you like to see in nTop?

401

17 Comments

Question for robotics engineers: What does this color map represent? Hint: the frame you see is what holds the sheet when industrial robots roboform it.

45

24 Comments

For those who want to know the answer: This is the map of the error between the two robots' end effectors after their out-of-the-box calibration. The wave pattern is interesting because the Cartesian movement of the end effector is a cumulation of the circular movements of the joints in a segmented robotic arm. Those circular movements of joints are showing up as wave patterns in the positional inaccuracy of the end effector. We use our own software and hardware stack to reduce this error and achieve accurate forming.

69

5 Comments

"Dynamical Analogies," authored by Harry F. Olson, is a groundbreaking work that examines the parallels between various physical systems through the use of analogies. The book delves into the mathematical and conceptual frameworks that facilitate the comparison of mechanical, electrical, acoustical, and other dynamic systems. #Olson presents the idea that different physical phenomena can be described using similar mathematical models, allowing for the transfer of concepts and solutions across different domains. The book is organized into sections that first cover the fundamental principles of dynamics, then apply these principles to various systems. Olson discusses the laws of motion, energy conservation, and the behavior of systems under different forces and constraints. He provides detailed examples and analogies, showing how electrical circuits can be compared to mechanical systems, and how acoustical properties can be related to electrical properties, among other comparisons. Throughout "#Dynamical #Analogies," Olson highlights the importance of understanding these relationships to solve complex problems in engineering and physics. Known for its clarity, comprehensive coverage, and practical approach, the book is a valuable resource for students and professionals in fields such as mechanical engineering, electrical engineering, acoustics, and applied physics. The insights offered in the book into the interconnectedness of dynamic systems have had a lasting impact on how these systems are analyzed and understood. I hereby attach a copy of first edition published in 1943.

30

1 Comment

Pole vault is one of the most dramatic track events. Before you invest in a pole, understand the mechanics with Simulink. https://spr.ly/6040l1MYm

1

🚀 Excited to kick off my #30DaysOfCocotb journey! 🚀 For the next 30 days, I'll be diving deep into the world of Cocotb, an open-source framework that has revolutionized hardware verification. Cocotb, short for coroutine-based co-simulation testbench, enables us to write testbenches in Python, combining the power of Python's extensive libraries with traditional hardware verification. Why Cocotb? - Python Integration: Leveraging Python's simplicity and extensive libraries, Cocotb allows for more flexible and powerful testbenches. - Co-Simulation: It enables co-simulation with various HDL simulators, making it versatile for different environments. - Coroutine-Based Testing: Utilizing Python's async capabilities, it offers a more intuitive way to handle concurrency in testbenches. Throughout this series, I'll share insights, examples, and tips on using Cocotb effectively for hardware verification. Whether you're a seasoned verification engineer or just starting out, I hope you'll find this journey informative and inspiring. Stay tuned for daily updates and feel free to join the conversation with your questions and experiences! #HardwareVerification #Cocotb #Python #Verification #VLSI #DigitalDesign #EDA

28

6 Comments

Can a Parasail Turn Cargo Ships into Sailboats? ⛵ Airseas is harnessing wind energy to propel massive ships with advanced systems designed using Simulink and Model-Based Design. MathWorks CEO Jack Little explains how this engineered system works and how Airseas may transform the cargo shipping industry 👇 #ModelBasedDesign #Simulink #Engineering Learn about other customers’ technical achievements using MATLAB and Simulink: https://spr.ly/6047SrtMz

1

Can a Parasail Turn Cargo Ships into Sailboats? ⛵ Airseas is harnessing wind energy to propel massive ships with advanced systems designed using Simulink and Model-Based Design. MathWorks CEO Jack Little explains how this engineered system works and how Airseas may transform the cargo shipping industry 👇 #ModelBasedDesign #Simulink #Engineering Learn about other customers’ technical achievements using MATLAB and Simulink: https://spr.ly/6041SrtpT

2