Navigation of tele-robot in dynamic environment using in-situ intelligence

A system, comprising:

• an input/output interface;
• one or more memories;
• and
• one or more hardware processors, the one or more memories coupled to the one or more hardware processors, wherein the one or more hardware processors are configured to execute programmed instructions stored in the one or more memories, to:
• initiate, via one or more hardware processors, a communication between the tele-robot and the master, where the communication is based on an application layer communication semantic, wherein the communication is associated with a plurality of modes of operation that includes a manual mode and an autonomous mode, and wherein the tele-robot is located at a remote location in the dynamic environment;
• receive, via one or more hardware processors, the communication regarding the mode of operation at the tele-robot from the master and determining whether a current mode of operation is the manual mode or the autonomous mode based on the application layer communication semantic and initiating a semi-persistent logical connection where a master node of the master becomes an observer observing the state of the tele-robot on the move and the tele-robot becomes a publisher/producer by periodically updating status;
• and
• perform, via one or more hardware processors, upon determining the plurality of modes of operation:
• on determining the current mode of operation of the tele-robot to be the manual mode, perform a manual navigation of the tele-robot, as commanded by the master based on a manual communication semantic;
• on determining the current mode of operation of the tele-robot to be the autonomous mode, perform the in-situ intelligent navigation of the tele-robot, comprising the steps of:
• receiving a depth map of the dynamic environment from a plurality of sensors and a goal point from the master, wherein a plurality of objects comprises of a plurality of static objects, a plurality of interactive objects and a plurality of dynamic objects, wherein an interactive object is an object which navigates or moves by application of an external force and a dynamic object navigates/or moves without the external force with a linear velocity (D) and an angular velocity (T), wherein the external force refers to a force external to the dynamic object or the interactive object;
• estimating a current position of the tele-robot based on a Simultaneous localization and mapping (SLAM) technique using a velocity associated with the tele-robot, wherein the velocity of the tele-robot is obtained using the plurality of sensors and the velocity includes an angular velocity and a linear velocity;
• generating a geo-centric top view map from the depth map using the current position of the tele-robot based on a spatial transform (ST) technique, wherein the depth map is converted to an ego-centric map which is a top view of the depth map and the ego-centric map is projected at the current position of the tele-robot using the ST technique to generate the geo-centric top view map, wherein the geo-centric top view map comprises the current position of the tele-robot and location details of the plurality of objects;
• receiving a video feed of the dynamic environment at the tele-robot using the plurality of sensors, wherein the video feed is in RGB format;
• identifying the plurality of interactive objects and the plurality of dynamic objects in the video feed based on an object detection technique;
• estimating, on determining an object of the plurality of objects as being an interactive object or a dynamic object:
• on determining the object is an interactive object, estimating an interactive object position of the interactive object in the geo-centric map based on a position estimation technique;
• and
• on determining the object is a dynamic object, estimating a future position of the dynamic object in the geo-centric map based on a future position estimation technique using a motion model, wherein the future position estimation technique includes the following steps:
• generating an ego centric projection of the dynamic object in the geo-centric map based on a depth segmentation technique, wherein the generation of the ego centric projection of the dynamic object in the geo-centric map based on the depth segmentation technique includes segregation of a depth map of the dynamic object from the rest of the depth map
• obtaining a geo centric representation of the dynamic object by passing an ego-centric map in a Spatial Transform (ST) using the velocity;
• and
• estimating the future position of the dynamic object based on the geo centric representation of the dynamic object using the motion model;
• and
• determining a navigation local policy for the tele-robot by passing the geo-centric map through a Fast Marching Motion Planner (FMMP) using the goal point, the current position of the tele-robot and the interactive object position or the future dynamic object position.

A non-transitory computer-readable medium having embodied thereon a computer readable program, wherein the computer readable program when executed by one or more hardware processors, cause:

• initiating, via one or more hardware processors, a communication between the tele-robot and the master, where the communication is based on an application layer communication semantic, wherein the communication is associated with a plurality of modes of operation that includes a manual mode and an autonomous mode, and wherein the tele-robot is located at a remote location in the dynamic environment;
• receiving, via one or more hardware processors, the communication regarding the mode of operation at the tele-robot from the master and determining whether a current mode of operation is a manual mode or an autonomous mode based on the application layer communication semantic and initiating a semi-persistent logical connection where a master node of the master becomes an observer observing the state of the tele-robot on the move and the tele-robot becomes a publisher/producer by periodically updating status;
• and
• performing, via one or more hardware processors, upon determining the plurality of modes of operation
• on determining the current mode of operation of the tele-robot to be the manual mode, performing a manual navigation of the tele-robot, as commanded by the master based on a manual communication semantic;
• on determining the current mode of operation of the tele-robot to be the autonomous mode, performing the in-situ intelligent navigation of the tele-robot, comprising the steps of:
• receiving a depth map of the dynamic environment from a plurality of sensors and a goal point from the master, wherein a plurality of objects comprises of a plurality of static objects, a plurality of interactive objects and a plurality of dynamic objects, wherein an interactive object is an object which navigates or moves by application of an external force and a dynamic object navigates or moves without the external force with a linear velocity (D) and an angular velocity (T), wherein the external force refers to a force external to the dynamic object or the interactive object;
• estimating a current position of the tele-robot based on a Simultaneous localization and mapping (SLAM) technique using a velocity associated with the tele-robot, wherein the velocity of the tele-robot is obtained using the plurality of sensors and the velocity includes an angular velocity and a linear velocity;
• generating a geo-centric top view map from the depth map using the current position of the tele-robot based on a spatial transform (ST) technique, wherein the depth map is converted to an ego-centric map which is a top view of the depth map and the ego-centric map is projected at the current position of the tele-robot using the ST technique to generate the geo-centric top view map, wherein the geo-centric top view map comprises the current position of the tele-robot and location details of the plurality of objects;
• receiving a video feed of the dynamic environment at the tele-robot using the plurality of sensors, wherein the video feed is in RGB format;
• identifying the plurality of interactive objects and the plurality of dynamic objects in the video feed based on an object detection technique;
• estimating, on determining an object of the plurality of objects as being an interactive object or a dynamic object:
• on determining the object is an interactive object, estimating an interactive object position of the interactive object in the geo-centric map based on a position estimation technique;
• and
• on determining the object is a dynamic object, estimating a future position of the dynamic object in the geo-centric map based on a future position estimation technique using a motion model, wherein the future position estimation technique includes the following steps:
• generating an ego centric projection of the dynamic object in the geo-centric map based on a depth segmentation technique, wherein the generation of the ego centric projection of the dynamic object in the geo-centric map based on the depth segmentation technique includes segregation of a depth map of the dynamic object from the rest of the depth map
• obtaining a geo centric representation of the dynamic object by passing an ego-centric map in a Spatial Transform (ST) using the velocity;
• and
• estimating the future position of the dynamic object based on the geo centric representation of the dynamic object using the motion model;
• and
• determining a navigation local policy for the tele-robot by passing the geo-centric map through a Fast Marching Motion Planner (FMMP) using the goal point, the current position of the tele-robot and the interactive object position or the future dynamic object position.