Robotics Research (United States)

Underwater vehicles have emerged as a critical technology for exploring aquatic environments. The deployment of multivehicle systems has gained substantial interest due to their capability to perform collaborative tasks with improved efficiency. However, achieving precise localization of a leader underwater vehicle within a multivehicle configuration remains a significant challenge, particularly in dynamic and complex underwater conditions. To address this issue, this article presents a novel trimodal sensor fusion neural network approach that integrates optical, acoustic, and pressure sensors to localize the leader vehicle. The proposed method leverages the unique strengths of each sensor modality to improve localization accuracy and robustness. Specifically, optical sensors provide high-resolution imaging for precise relative positioning, acoustic sensors enable long-range detection and ranging, and pressure sensors offer environmental context awareness. The fusion of these sensor modalities is implemented using a deep learning architecture designed to extract and combine complementary features from raw sensor data. The effectiveness of the proposed method is validated through a custom-designed testing platform and field test. Extensive data collection and experimental evaluations demonstrate that the trimodal approach significantly improves the accuracy and robustness of leader localization, outperforming both single-modal and dual-modal methods.

Artificial muscle-driven modular soft robots exhibit significant potential for executing complex tasks. However, their broader applicability remains constrained by the lack of dynamic model-based control strategies tailored for multidegree-of-freedom (DOF) configurations. This article presents a novel design of a 2-DOF robotic wrist, envisioned as a fundamental building block for such advanced robotic systems. The wrist module is actuated by twisted and coiled actuators (TCAs) and utilizes a compact 3RRRR parallel mechanism to achieve a lightweight structure with enhanced motion capability. A comprehensive Lagrangian dynamic model is developed to capture the module’s complex nonlinear behavior. Leveraging this model, a nonlinear model predictive controller (NMPC) is designed to ensure accurate trajectory tracking. A physical prototype of the robotic wrist is fabricated, and extensive experiments are performed to validate its motion performance and the fidelity of the proposed dynamic model. Subsequently, comparative evaluations between the NMPC and a conventional proportional–integral–derivative controller are conducted under various operating conditions. Experimental results demonstrate the effectiveness and robustness of the dynamic model-based control approach in managing the motion of TCA-driven robotic wrists. Finally, to illustrate its practical utility and integrability, the wrist module is incorporated into a multisegment soft robotic arm, where it successfully executes a trajectory tracking task.

In modern warehouse automation, mobile robots are essential for enhancing operational efficiency by autonomously navigating to pick and transport items. Effective path planning is crucial for these robots to move through complex environments, avoid obstacles, and minimize travel time. This study evaluates two prominent path planning algorithms Dijkstra and A*—implemented on a mobile robot within a simulated 3D warehouse environment using CoppeliaSim. Three distinct rack locations were analyzed to assess the performance of both algorithms concerning path optimality, computational efficiency, and real-time applicability. Simulation results indicate that while both algorithms successfully generated safe and accurate paths, A* outperformed Dijkstra in terms of speed and path efficiency. A*'s heuristic-driven approach resulted in lower computational load and faster execution time, making it more suitable for real-time warehouse operations where responsiveness is critical. The insights gained provide valuable guidance for robotics engineers and developers in selecting appropriate path planning strategies for autonomous navigation in industrial settings.

Abstract Continuous invasive brain–computer interfaces (BCIs) translate neural activity into continuous control signals. During ongoing control, fluctuations in these signals can reflect either transient execution noise or genuine changes in user intent, yet most BCI control systems do not explicitly distinguish between these possibilities. Assistive controllers must therefore determine whether variability should be stabilized as noise or expressed as intentional changes in movement. Here we evaluate a confidence-modulated shared-control framework that adaptively integrates decoded neural commands with a temporal prior to balance stabilization and responsiveness. Using macaque BCI navigation tasks that impose opposing control demands, we show that shared control nearly eliminates execution failures in obstacle-avoidance tasks while preserving the directional structure of commands. When goals change abruptly, however, the same temporal prior introduces transient inertia. Resetting the prior restores baseline performance, revealing a fundamental stabilization–responsiveness trade-off imposed by temporal priors during continuous arbitration.