Inventec had no grants in robotics as a theme in Q2 2024.

Recent Patents

Application: Transition method of locomotion gait of robot (Patent ID: US20240157548A1)

The patent filed by Inventec Corp. describes a transition method for the locomotion gait of a robot, involving the deployment of neural network-based gait policies and a transition-net to facilitate smooth transitions between different locomotion patterns. The method includes selecting source and destination policies, simulating transitions, recording results in a database, training the transition-net, and using a meta-controller to determine when to execute queued policies based on success probabilities.

Additionally, the method involves collecting real-world locomotion data, creating neural network models for different gaits, training them using reinforcement learning with domain randomization, and storing the models in a library upon completion. The neural network models consist of input layers representing joint angles, robot direction, velocities, and ground contact, middle layers serving as latent states, and output layers for proportional-derivative control parameters, with the transition-net being a feed-forward neural network with specific layer configurations and activation functions. The simulation of transition operations utilizes the Issac Gym simulator, showcasing a comprehensive approach to enhancing robot locomotion capabilities through intelligent policy transitions.

To know more about GlobalData’s detailed insights on Inventec, buy the report here.

Data Insights

From

The gold standard of business intelligence.

Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.

Be better informed