English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this paper, we present a learning-based approach that allows a robot toquickly follow a reference path defined in joint space without exceeding limitson the position, velocity, acceleration and jerk of each robot joint. Contraryto offline methods for time-optimal path parameterization, the reference pathcan be changed during motion execution. In addition, our approach can utilizesensory feedback, for instance, to follow a reference path with a bipedal robotwithout losing balance. With our method, the robot is controlled by a neuralnetwork that is trained via reinforcement learning using data generated by aphysics simulator. From a mathematical perspective, the problem of tracking areference path in a time-optimized manner is formalized as a Markov decisionprocess. Each state includes a fixed number of waypoints specifying the nextpart of the reference path. The action space is designed in such a way that allresulting motions comply with the specified kinematic joint limits. The rewardfunction finally reflects the trade-off between the execution time, thedeviation from the desired reference path and optional additional objectiveslike balancing. We evaluate our approach with and without additional objectivesand show that time-optimized path tracking can be successfully learned for bothindustrial and humanoid robots. In addition, we demonstrate that networkstrained in simulation can be successfully transferred to a real robot.

Learning Time-optimized Path Tracking with or without Sensory Feedback