In this paper, we propose a novel design of an electrically actuated robotic leg, called the DecARt (Decoupled Actuation Robot) Leg, aimed at performing agile locomotion. This design incorporates several new features, such as the use of a quasi-telescopic kinematic structure with rotational motors for decoupled actuation, a near-anthropomorphic leg appearance with a forward facing knee, and a novel multi-bar system for ankle torque transmission from motors placed above the knee. To analyze the agile locomotion capabilities of the design numerically, we propose a new descriptive metric, called the "Fastest Achievable Swing Time" (FAST), and perform a quantitative evaluation of the proposed design and compare it with other designs. Then we evaluate the performance of the DecARt Leg-based robot via extensive simulation and preliminary hardware experiments.

Accurate system identification is crucial for reducing trajectory drift in bipedal locomotion, particularly in reinforcement learning and model-based control. In this paper, we present a novel control framework that integrates system identification into the reinforcement learning training loop using differentiable simulation. Unlike traditional approaches that rely on direct torque measurements, our method estimates system parameters using only trajectory data (positions, velocities) and control inputs. We leverage the differentiable simulator MuJoCo-XLA to optimize system parameters, ensuring that simulated robot behavior closely aligns with real-world motion. This framework enables scalable and flexible parameter optimization. It supports fundamental physical properties such as mass and inertia. Additionally, it handles complex system nonlinear behaviors, including advanced friction models, through neural network approximations. Experimental results show that our framework significantly improves trajectory following. It reduces rotational deviation by 75% and increases travel distance in the commanded direction by 46% compared to a baseline reinforcement learning method.

This paper proposes a model of a distributed system that sorts items using robotic agents. The system operates with multiple robotic agents that work collaboratively to sort items in accordance with their destinations, and a controller manages the behavior and states of the agents. The distributed system is designed to be scalable and fault-tolerant, with the ability to handle large volumes of items. Scalability is provided by a changeable number of robotic agents presented. The distributed sorting system offers flexibility in the allocation of resources and completes tasks more efficiently than traditional conveyor sorting systems do. In this paper we present the robotic agents fleet control system structure, the agents task modeling system and the sorting center space modeling system. Each agent receives an action schedule to execute from the control system and implements movements in the sorting center space modeling system. The fleet management system calculates the movement schedule for each agent, taking into account the movement, the tasks of all robotic agents and the states of the sorting center components. With the help of the distributed system, the sorting center throughput is estimated for the given model parameters. A series of experiments is conducted to provide this estimation.

We propose a method of improving detection precision (mAP) with the help of the prior knowledge about the scene geometry: we assume the scene to be a plane with objects placed on it. We focus our attention on autonomous robots, so given the robot's dimensions and the inclination angles of the camera, it is possible to predict the spatial scale for each pixel of the input frame. With slightly modified YOLOv3-tiny we demonstrate that the detection supplemented by the scale channel, further referred as S, outperforms standard RGB-based detection with small computational overhead.