Systems, devices, and methods for grasping by multi-purpose robots
Systems, devices, and methods for training and operating (semi-)autonomous robots to complete multiple different work objectives are described. A robot control system stores a library of reusable work primitives each corresponding to a respective basic sub-task or sub-action that the robot is operative to autonomously perform. A work objective is analyzed to determine a sequence (i.e., a combination and/or permutation) of reusable work primitives that, when executed by the robot, will complete the work objective. The robot executes the sequence of reusable work primitives to complete the work objective. The reusable work primitives may include one or more reusable grasp primitives that enable(s) a robot's end effector to grasp objects. Simulated instances of real physical robots may be trained in simulated environments to develop control instructions that, once uploaded to the real physical robots, enable such real physical robots to autonomously perform reusable work primitives.
1. A method of operation of a robot to grasp an object, wherein the robot includes:
a robotic hand having multiple fingers and an opposable thumb;
at least one processor operative to control actuations of the robotic hand;
at least one sensor communicatively coupled to the at least one processor; and
a non-transitory processor-readable storage medium communicatively coupled to the at least one processor, the non-transitory processor-readable storage medium storing a library of reusable grasp primitives and processor-executable instructions that, when executed by the at least one processor, cause the robotic hand to autonomously perform the reusable grasp primitives in the library of reusable grasp primitives, the method comprising:
collecting data about the object by the at least one sensor;
analyzing the data by the at least one processor to determine a geometry of the object;
simulating, by the at least one processor, the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine a particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives; and
executing, by the robotic hand, the particular reusable grasp primitive to grasp the object.
2. The method of claim 1 wherein executing, by the robotic hand, the particular reusable grasp primitive to grasp the object includes executing, by the at least one processor, processor-executable instructions that cause the robotic hand to autonomously perform the particular reusable grasp primitive to grasp the object.
3. The method of claim 1 wherein the at least one sensor includes at least one optical sensor, and wherein:
collecting data about the object by the at least one sensor includes collecting optical data about the object by the at least one optical sensor; and
analyzing the data by the at least one processor to determine a geometry of the object includes analyzing the optical data by the at least one processor to determine the geometry of the object.
4. The method of claim 1 , further comprising:
analyzing additional data about the object by the at least one processor to determine at least one additional parameter of the object, the method further comprising selecting, by the at least one processor, the particular reusable grasp primitive from the library of reusable grasp primitives based, at least in part, on both: i) simulating, by the at least one processor, the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine the particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives, and ii) the at least one additional parameter of the object.
5. The method of claim 4 , further comprising:
collecting the additional data about the object by the at least one sensor, and wherein the additional data about the object is selected from a group consisting of: a hardness of the object, a rigidity of the object, an identity of the object, a function of the object, and a mass of the object.
6. The method of claim 4 wherein the robot further includes at least one receiver communicatively coupled to the at least one processor, and wherein the method further comprises:
receiving the additional data about the object by the receiver.
7. The method of claim 1 wherein the non-transitory processor-readable storage medium further stores data about a work objective to be performed by the robot, the work objective involving grasping the object, and the method further comprising -selecting, by the at least one processor, the particular reusable grasp primitive from the library of reusable grasp primitives based, at least in part, on both: i) simulating, by the at least one processor, the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine the particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives, and ii) the data about the work objective to be performed by the robot.
8. A robot comprising:
a body;
a robotic hand mechanically coupled to the body, the robotic hand having multiple fingers and an opposable thumb;
at least one processor operative to control actuations of the robotic hand;
at least one sensor communicatively coupled to the at least one processor; and
a non-transitory processor-readable storage medium communicatively coupled to the at least one processor, the non-transitory processor-readable storage medium storing a library of reusable grasp primitives and processor-executable instructions that, when executed by the at least one processor, cause the robot to:
collect data about an object;
analyze the data to determine a geometry of the object;
simulate the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine a particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives; and
execute the particular reusable grasp primitive to grasp the object.
9. The robot of claim 8 wherein the processor-executable instructions that, when executed by the at least one processor, cause the robot to execute the particular reusable grasp primitive to grasp the object, cause the robotic hand to autonomously perform the particular reusable grasp primitive to grasp the object.
10. The robot of claim 8 wherein the at least one sensor includes at least one optical sensor, and wherein the processor-executable instructions that, when executed by the at least one processor, cause the robot to collect data about the object, cause the at least one optical sensor to collect optical data about the object.
11. The robot of claim 8 wherein the non-transitory processor-readable storage medium further stores processor-executable instructions that, when executed by the at least one processor, cause the robot to:
analyze additional data about the object to determine at least one additional parameter of the object; and
select a particular reusable grasp primitive from the library of reusable grasp primitives based, at least in part, on both: i) simulating the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine a particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives, and ii) the at least one additional parameter of the object.
12. The robot of claim 11 wherein the non-transitory processor-readable storage medium further stores processor-executable instructions that, when executed by the at least one processor, cause the robot to collect the additional data about the object, and wherein the additional data about the object is selected from a group consisting of: a hardness of the object, a rigidity of the object, an identity of the object, a function of the object, and a mass of the object.
13. The robot of claim 11 , further comprising:
at least one receiver communicatively coupled to the at least one processor, and wherein the non-transitory processor-readable storage medium further stores processor-executable instructions that, when executed by the at least one processor, cause the robot to receive the additional data about the object.
14. The robot of claim 8 wherein the non-transitory processor-readable storage medium further stores data about a work objective to be performed by the robot, the work objective involving grasping the object, and processor-executable instructions that, when executed by the at least one processor, cause the robot to select the particular reusable grasp primitive from the library of reusable grasp primitives based, at least in part, on both: i) simulating the robotic hand grasping the object with each reusable grasp primitive in the library of reusable grasp primitives to determine a particular reusable grasp primitive in the library of reusable grasp primitives that best fits the geometry of the object relative to all other grasp primitives in the library of reusable grasp primitives, and ii) the data about the work objective to be performed by the robot.