Robot systems incorporating cloud services systems
Robot systems including a robot, an on-premise computing device, and a cloud services system are disclosed. An on-premise computing device includes a processor and a non-transitory memory device storing machine-readable instructions that, when executed by the processor, cause the processor to receive raw sensor data from a robot, pre-process the sensor data to remove unnecessary information from the sensor data and obtain pre-processed data, transmit the pre-processed data to a cloud services system, receive, from the cloud services system, an object type, an object pose, and an object location of an object corresponding to the pre-processed data, and transmit a signal to the robot based on at least one of the object type, the object pose, and the object location of the object.
1. An on-premise computing device comprising a processor and a non-transitory memory device storing machine-readable instructions that, when executed by the processor, cause the processor to:
receive raw sensor data from a robot;
pre-process the raw sensor data to remove unnecessary information from the raw sensor data and obtain pre-processed data;
transmit the pre-processed data to a cloud services system;
receive, from the cloud services system, an object type, an object pose, and an object location of an object corresponding to the pre-processed data; and
transmit a signal to the robot based on at least one of the object type, the object pose, and the object location of the object, the signal comprising one or more instructions usable by the robot to manipulate the object.
2. The on-premise computing device of claim 1 , wherein a size of the pre-processed data is less than a size of the raw sensor data.
3. The on-premise computing device of claim 1 , wherein the machine-readable instructions further cause the processor to receive path planning instructions from the cloud services system and transmit the path planning instructions to the robot.
4. The on-premise computing device of claim 1 , wherein the machine-readable instructions that cause the processor to receive the raw sensor data from the robot further cause the processor to receive the raw sensor data from a co-located robot.
5. The on-premise computing device of claim 1 , wherein the machine-readable instructions further cause the processor to receive a command via a user interface prior to receiving the raw sensor data from the robot, wherein the command corresponds to an instruction for the robot to perform a task.
6. The on-premise computing device of claim 1 , wherein the raw sensor data comprises data corresponding to the object and a region detection bubble around the object.
7. The on-premise computing device of claim 1 , wherein the machine-readable instructions further cause the processor to authenticate a user prior to receiving the raw sensor data from the robot, wherein the user is authenticated by comparing an input from the user that corresponds to known objects in an area adjacent to the robot with stored information regarding the known objects.
8. The on-premise computing device of claim 1 , wherein the object type is derived from tagged object data corresponding to objects that have previously been tagged by a user.
9. A cloud services system comprising a processor and a non-transitory memory device storing machine-readable instructions that, when executed by the processor, cause the processor to:
receive pre-processed data from an on-premise computing device, wherein the pre-processed data is data that has been extracted by the on-premise computing device from raw sensor data that originated from a robot;
determine an object type, an object pose, and an object location of an object from the pre-processed data;
transmit the object type, the object pose, and the object location to the on-premise computing device;
generate instructions for causing the robot to manipulate the object; and
transmit the instructions to the on-premise computing device.
10. The cloud services system of claim 9 , wherein the pre-processed data is image data and the object type and the object pose are determined by applying an object recognition algorithm to the image data.
11. The cloud services system of claim 9 , wherein the machine-readable instructions further cause the processor to perform a de-duplication process that evaluates a plurality of historical measurements of an object to determine if a current object is:
a same object as one or more historical measurements of the plurality of historical measurements, or
a second object that is different from the object.
12. The cloud services system of claim 9 , wherein the machine-readable instructions further cause the processor to:
generate a robot path plan, and
transmit the robot path plan to the on-premise computing device.
13. The cloud services system of claim 9 , wherein the machine-readable instructions further cause the processor to authenticate a user prior to receiving the pre-processed data from the on-premise computing device, wherein the user is authenticated by comparing an input from the user that corresponds to known objects in an area adjacent to the robot with stored information regarding the known objects.
14. The cloud services system of claim 9 , wherein the object type is determined from tagged object data corresponding to objects that have previously been tagged by a user.
15. A method of controlling a robot, the method comprising:
receiving raw sensor data from the robot at an on-premise computing device;
reducing, by the on-premise computing device, a size of the raw sensor data to generate processed sensor data;
transmitting, by the on-premise computing device, the processed sensor data to a cloud services system;
receiving, by the on-premise computing device, an object type, an object pose, and an object location from the cloud services system, wherein the object type, the object pose, and the object location is determined by the cloud services system from the processed sensor data; and
transmitting, by the on-premise computing device, a signal to the robot based on at least one of the object type, the object pose, and the object location of the object, the signal comprising one or more instructions usable by the robot to manipulate the object.
16. The method of claim 15 , wherein the object type and the object pose are determined by an object recognition algorithm.
17. The method of claim 15 , wherein receiving the object type, the object pose, and the object location comprises receiving information about the object that is generated as a result of a de-duplication process that evaluates a plurality of historical measurements of the object to determine if a current object is a same object as one or more historical measurements of the plurality of historical measurements or a different object.
18. The method of claim 15 , further comprising receiving a command via a user interface prior to receiving the raw sensor data from the robot, wherein the command corresponds to an instruction for the robot to perform a task.
19. The on-premise computing device of claim 1 , wherein the pre-processed data is image data and the object type and the object pose are determined by the cloud services system applying an object recognition algorithm to the image data.
20. The method of claim 15 , further comprising:
receiving, by the on-premise computing device, path planning instructions from the cloud services system; and
transmitting the path planning instructions to the robot.