Rapyuta Robotics Co., Ltd.

Hendrik Skubch

Wen Zheng Li

Embodiments herein disclose methods and systems to avoid congestion in an area by a plurality of robots awaiting access to a common resource. A current robot action and a sensor data of one of the plurality of robots is received at a cloud node. Based on the received current robot action and the sensor data, a determination is made whether the one of the plurality of robots has to access the common resource. Next, a spatial query is executed at the cloud node to determine a location in the area for positioning the one of the plurality of robots awaiting access to the common resource based on a static spatial condition that requires a congestion-free movement when the robot is positioned at the location in the area. Next, based on a collaboratively determined priority order, the common resource is accessed by the one of the plurality of robots.

Congestion avoidance and common resource access management for multiple robots

Embodiments herein disclose methods and systems for generating a 2D-navigation map for collision-free navigation by a plurality of robots. Initially, a 2D-navigation map generator executing at a cloud map server initiates a particle simulation to add a plurality of bubbles to an image of an area. Next, the 2D-navigation map generator executing at the cloud map server, processes the bubble simulation to determine a set of bubbles from the plurality of robots on the image that are overlapping free spaces of the area in the image. Next, the 2D-navigation map generator executing at the cloud map server, generates the 2D-navigation map by connecting the set of bubbles using a plurality of edges, wherein the plurality of edges indicate a navigation path to navigate within the area. Finally, based on the generated 2D-navigation map, the plurality of robots traverses from a current location to another location.

Generating a 2D-navigation map for collision-free navigation by multiple robots

A system and method to control execution of a centralized and decentralized controlled plan, has been described. A plan execution engine executing at a cloud node, receives sensor data captured by one or more sensors at a plurality of autonomous robots and a plan execution status of the centralized controlled plan. The plan execution engine executing at the cloud node, determines whether the plurality of autonomous robots satisfy a transition condition. Next a determination is made one or more activated constraints and task allocation for one or more autonomous robots in the next state. Next the plan execution engine executing at the cloud node and autonomous robots collaboratively determine a constraint solution for the activated one or more plan constraints. Finally based on the determined constraint solution, the one or more plan execution engines sends instructions to an actuator for executing the task included in the centralized controlled plan.

Growing community of inventors

Tokyo, Japan

Hendrik Skubch