Algoryx’s Post

Excited to be part of this cutting-edge R&D collaboration, where I get to do what I’m most passionate about: Merging AI and Physics to drive machine autonomy! It’s amazing to see how the gap between fundamental science and real-world engineering has become so narrow, allowing us to bring innovation to life faster than ever before. (Read more in the link provided!)

Building a sustainable future includes creating machines that minimize environmental impact. That includes manifacturing but also how the machine operates. Being able to simulate a machine in a virtual environment while having access to energy consumption, wear and tear will help engineers to design algoritms for control that can have a substantial effect on the environment. ”What if” scenarios is now available on the earth or on the moon (or other celestial bodies)!

🌟 Proud to share that our team secured second place in the Leiden University Artificial Intelligence and Robotics competition! 🥈 Collaborating with Adam Bujna, Tarik Začiragić, Juan Mediavilla Sáez, Alexander Cremer, and Xocas Álvarez was an incredible experience. The competition was a thrilling challenge that involved developing Artificial Intelligence from scratch for robots to play tag in a complex arena. Our approach utilized a hierarchical layered behavior model. In this model, the lower layers managed fundamental, high-priority tasks, such as collision avoidance, ensuring the robot navigated safely. The higher layers handled more complex tasks, including following or avoiding other robots. This structure allowed for a robust and adaptable system capable of effectively responding to a dynamic environment. For the hiding role, our robot employed a sophisticated force field mechanism in the higher layers to avoid approaching robots. This mechanism was mathematically modeled by summing the repulsive forces exerted by each nearby robot detected by the LiDAR. In a very simplified form, the equation is:   n F=∑ k (r−r_i)/∥r−r_i∥^3   i=1 For the seeker role, we delved deep into computer vision, converting inputs from the RGB and depth cameras into processable data. This data was then filtered to train a convolutional neural network, successfully detecting the other robots. Once detected, our robot would switch from the exploration algorithm to the tracking algorithm, effectively "tagging" its rivals. Beyond the technical triumphs, this project was a testament to the power of teamwork and the vital role of soft skills. Effective communication, strategic planning, and task delegation were as crucial as the coding and algorithms. 🚀🤖 #AI #Teamwork #Innovation #ComputerVision #UniversiteitLeiden #SoftSkills #LeidenAIRCompetition

🚀 Exploring the Future of Distributed Systems with Swarm Intelligence 🐝 After an enlightening discussion with my good friend Daniel Mathieson, we dove deep into the fascinating world of swarm intelligence and how it shapes modern distributed systems. This led me to explore multi-agent systems (MAS) on hyperbolic graphs—a powerful framework for modelling everything from autonomous vehicle fleets to global logistics networks. In my latest article, I explore how agents navigate these complex, hierarchical environments, balancing decision-making based on centrality, avoiding overuse of resources, and leveraging Topological Data Analysis (TDA) to optimize their paths. 💡 Whether you're working in smart cities, robotics, or network optimization, this deep dive into MAS will offer insights into the potential of decentralized, intelligent agents to solve complex, distributed problems. Check it out here #SwarmIntelligence #MultiAgentSystems #DistributedSystems #HyperbolicGraphs #AI #DataScience

ATOM P1: Advancing Humanoid Robotics with Innovative Neural Network Design（2024） Foxtech Robot #HumanoidRobotics #ReinforcementLearning #RobotControl Introducing the ATOM P1, an entry-level bipedal robot designed for cutting-edge research in humanoid robot motion control algorithms and reinforcement learning. Neural Network Architecture: The advanced architecture of the ATOM P1 is engineered to handle significant interference while generating adaptive control policies. This flexibility allows the same network to be utilized across various robots and scenarios, enhancing its adaptability in diverse environments. Data Generation Mechanism: Our iterative pre-training methodology efficiently produces and collects valid data, essential for developing high-performance policies that drive the robot's capabilities. Discover how the ATOM P1 is pushing the boundaries of humanoid robotics and paving the way for the future of intelligent automation. Don't miss out on the journey toward smarter robots! #HumanoidRobotics #ReinforcementLearning #RobotControl #ATOM_P1 #ArtificialIntelligence #InnovationInRobotics Discover More: https://lnkd.in/gMzYnesk... https://lnkd.in/gGsbSH3N

🎉 Excited to share that the work we have been working on for the past two years got accepted to #ECCV2024, the European Conference on Computer Vision. Our work #JDT3D (joint 3D object detection and tracking) aims to address the gaps in the current #Transformer based end-to-end detection and tracking architecture for #autonomousdriving. When I started my master, I strongly believed that object detection shouldn’t be done just on single frames, but should leverage the past frames of detection as much as possible while performing object tracking jointly in an implicit way. In this work, we took inspiration from the #Transformer (the now famous architecture in natural language processing), adopted it to 3D #LiDAR-based computer vision, and proposed several key optimization and training techniques (track sampling augmentation, confidence-based query propagation) to bridge the performance gap with the traditional two-step data association approaches. This work wouldn’t have been possible without the diligence and preseverance of Brian Cheong who is still working on further improvements in this area, and the unwavering support and supervision from Professor Steven Lake Waslander. As a commitment to the research community, we are preparing to open-source our work, so stay tuned by following us at Toronto Robotics and Artificial Intelligence Laboratory. Check out the full paper in detaills linked below. Paper arxiv link: https://lnkd.in/eVEBWU7S #autonomousdriving #computervision #lidar #objectdetection #multiobjecttracking #robotics #artificialintelligence

🚀 Exciting Breakthrough in LiDAR-Based Joint Detection and Tracking! 🚀 We're thrilled to share our latest research from the University of Toronto that has been accepted to #ECCV2024: "JDT3D: Addressing the Gaps in LiDAR-Based Tracking-by-Attention" by Brian Cheong, Jiachen (Jason) Zhou and Steven Lake Waslander. In computer vision, approaches trained in an end-to-end manner have been shown to perform better than traditional pipeline-based methods. However, within LiDAR-based object tracking, tracking-by-detection continues to achieve state-of-the-art performance without learning both tasks jointly. In our work, we explore the potential reasons for this gap and propose techniques to leverage the advantages of joint detection and tracking. 🌟 Key Highlights: - Innovative Approach: We propose JDT3D, a novel LiDAR-based joint detection and tracking model that leverages transformer-based decoders to propagate object queries over time, implicitly performing object tracking without an association step at inference. - Enhanced Techniques: We introduce track sampling augmentation and confidence-based query propagation to bridge the performance gap between tracking-by-detection (TBD) and joint detection and tracking (JDT) methods. - Real-World Impact: Our model is trained and evaluated on the nuScenes dataset, showcasing significant improvements in tracking accuracy and robustness. Check out the full paper linked below and join us at #ECCV2024! Paper: https://lnkd.in/gbT4EStA Code: https://lnkd.in/gU385pcW #autonomousdriving #computervision #tracking #objecttracking #robotics #3DVision #transformers #deeplearning

Could artificial intelligence (AI) and hyperspectral imaging (HSI) be combined to enhance environmental protection? Absolutely! According to HSI robotics, this integration will help support advancements in environmental monitoring, disaster response, agriculture, and mining. Plus, this convergence will help ensure robots perceive and analyze their surroundings with greater accuracy. Some notable features of HSI that facilitate these benefits include spectral resolution, non-invasive monitoring, and material identification. See more here: https://heyor.ca/wOT6LB Here at AvaWatz, we believe that technological advancements have the potential to revolutionize environmental protection...take hyperspectral imaging and AI for example! #AI #HSI #EnvironmentalProtection

𝐀𝐝𝐯𝐚𝐧𝐜𝐢𝐧𝐠 𝐂𝐡𝐚𝐫𝐚𝐜𝐭𝐞𝐫 𝐂𝐨𝐧𝐭𝐫𝐨𝐥 𝐟𝐨𝐫 𝐑𝐞𝐚𝐥-𝐖𝐨𝐫𝐥𝐝 𝐑𝐨𝐛𝐨𝐭𝐬 Recent breakthroughs in physics-based character control are helping us learn from unstructured motion data. However, it’s still a challenge to create a single control policy that works with diverse motions and can be used on real-world robots. In a new approach, researchers developed a two-stage method to address this: Stage 1: Train a variational autoencoder (VAE) to extract a latent space encoding from unstructured motion data, capturing complex movement patterns. Stage 2: Use the learned latent code to train a policy that maps kinematic input to dynamics-aware output, enabling accurate motion imitation. This separation allows the model to learn better, avoid errors, and adapt to unseen motions. The method has been tested in simulations and on bipedal robots, successfully bringing complex, dynamic movements into the real world. Please check reference for my post in comment section #Robotics #AI #PhysicsBasedControl #MachineLearning #RobotMotion #ImitationLearning #CharacterControl

Robot dogs mimic natural human oscillation patterns to run more efficiently. https://buff.ly/40Vxwh3