Dataset achieve significant performance improvements in both closed-loop and open-loop metrics.

An open research question in robotics is how to combine the benefits of model-free reinforcement learning (RL) - known for its strong task performance and flexibility in optimizing general reward formulations - with the robustness and online replanning capabilities of model predictive control (MPC). This paper provides an answer by introducing a new framework called Actor-Critic Model Predictive Control. The key idea is to embed a differentiable MPC within an actor-critic RL framework. The proposed approach leverages the short-term predictive optimization capabilities of MPC with the exploratory and end-to-end training properties of RL. The resulting policy effectively manages both short-term decisions through the MPC-based actor and long-term prediction via the critic network, unifying the benefits of both model-based control and end-to-end learning. We validate our method in both simulation and the real world with a quadcopter platform across various high-level tasks. We show that the proposed architecture can achieve real-time control performance, learn complex behaviors via trial and error, and retain the robustness inherent to MPC.

MADISON, Wis., April 17, 2023 (SEND2PRESS NEWSWIRE) -- David P. Perrodin, Ph.D., collaborates with renowned actor Ben Hauck and audio streaming company Spotify to create the audio version of his book "The Velocity of Information – Human Thinking During Chaotic Times." Utilizing personal interviews and explorations of historical and contemporary events, "The Velocity of Information" gives you an expectation of how both individuals and the masses will behave during different stages of chaos. "This book provides twelve strategies to become more resilient to burnout and increase your survival chances during uncertain times," explains Dr. Perrodin. Released in 2022 in both print and eBook formats, Rowman & Littlefield published The Velocity of Information, which is now featured in distinguished libraries across the globe, including: The Library of Congress, Bavarian State Library in Munich, and The Chinese University of Hong Kong Library. Professional actor and narrator Ben Hauck jumped at the opportunity to narrate the audiobook.

To be useful tools in real scenarios, humanoid robots must realize tasks dynamically. This means that they must be capable of applying substantial forces, rapidly swinging their limbs, and also mitigating impacts that may occur during the motion. Towards creating capable humanoids, this letter presents the leg of the robot TELLO and demonstrates how it embodies two new fundamental design concepts for dynamic legged robots. The limbs follows the principles of: (i) Cooperative Actuation (CA), by combining motors in differential configurations to increase the force capability of the limb. We demonstrate that the CA configuration requires half the motor torque to perform a jump in comparison to conventional serial design configurations. And (ii) proximal actuation, by placing heavy motors near the body to reduce the inertia of the limb. To quantify the effect of motor placement on the robot's dynamics, we introduce a novel metric entitle Centroidal Inertia Isotropy (CII). We show that the design of state-of-the-art dynamic legged robots empirically increase the CII to improve agility and facilitate model-based control. We hope this metric will enable a quantifiable way to design these machines in the future.

But a team of researchers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and ETH Zurich hope to make drone cinematography more accessible, simple, and reliable. Then, on the fly, it generates control signals for a camera-equipped autonomous drone, which preserve that framing as the actors move. With our solution, if the subject turns 180 degrees, our drones are able to circle around and keep focus on the face. The researchers tested the system at CSAIL's motion-capture studio, using a quadrotor (four-propeller) drone.