3results about How to "Natural movement" patented technology

The application relates to the technical field of printed circuit board hot-press forming equipment, and particularly discloses a layered positioning and pressing device for high-density interconnection printed circuit boards, which comprises a rack, a pressing rack, a pressing mechanism, positioning pins, a floating assembly, a pneumatic mechanism and a guiding mechanism. The pneumatic mechanism comprises air holes, a plugging plug and an air pipe. The plugging plug closes the air holes during pressing, forms a negative pressure space after cooling and pressurized gas is output during resetting. The application can adsorb the circuit board for a short time at the beginning of mold opening, avoids the circuit board from being taken up with the pressing mechanism, and separates the circuit board through positive pressure flexible assistance. Meanwhile, the positioning pins can be lowered to avoid interference and damage during plate taking, and the pressing stability is improved.

This invention discloses a training method and system for quadruped robots based on deep reinforcement learning. The AMP training framework significantly enhances the robot's adaptability to complex environments. By distinguishing the robot's current motion sample distribution from that of a reference sample, the training aims to maximize the probability of misjudgment and the reward for reinforcement learning, enabling the robot to learn motion patterns that meet actual needs. When the environment changes, the robot can adjust its strategy promptly based on new information without remodeling, solving the problem of traditional methods being sensitive to environmental changes. Furthermore, this framework introduces imitation learning into reinforcement learning, providing effective guidance for robot learning, reducing the blindness of random exploration, accelerating the learning speed of effective strategies, and improving training efficiency. Simultaneously, the integrated hierarchical LSTM Actor network can extract and aggregate motion features at multiple time scales, allowing the robot to generate actions that consider both short-term gait and long-term trends, resulting in more natural and fluid movements. It also fully utilizes existing sample information, improving sample utilization and reducing training costs.

The invention discloses a skeletal animation interactive display method, and relates to the technical field of skeletal display, the method is applied to an interactive display terminal, and the method comprises the following steps: the interactive display terminal generates a standardized hierarchical joint tree based on a multi-view image or a draft, distributes a skin weight by adopting a conflict area dynamic isolation and weight self-optimization algorithm, and displays a skeletal animation. Constructing a joint motion parameterized linkage rule base, and forming an adjustable skeleton skin basic structure; the interactive display terminal integrates physical motion capture data, text instruction analysis data and user operation data, synchronously drives a joint tree through a multi-source data space-time collaborative fusion engine, drives the joint tree through combination of multi-source data space-time collaborative fusion and scene adaptive weight distribution, breaks through the limitation of traditional single data driving, and improves the user experience. Space-time synchronization of three types of data of physical motion capture, text instruction analysis and user operation is achieved, then the weight is dynamically adjusted according to the scene, the scene is rapidly moved to preferentially guarantee coherence, and the scene is finely adjusted to focus on accuracy.