| CPC G05B 19/4155 (2013.01) [B25J 9/163 (2013.01); B25J 13/085 (2013.01); G06N 20/00 (2019.01); G06Q 10/06395 (2013.01); G06Q 10/087 (2013.01); G05B 2219/40269 (2013.01)] | 20 Claims |

|
1. A robotic system, comprising:
an end effector configured to manipulate an item;
a robotic arm attached with the end effector and configured to move the end effector;
a set of sensors configured to generate multi-dimensional time-series data corresponding to at least one of: a manipulation of the item by the end effector or a movement of the end effector by the robotic arm; and
a computer system comprising one or more processors and one or more memories, the one or more memories storing instructions that, upon execution by the one or more processors, configure the computer system to:
control an inventory action associated with the item by at least controlling an operation of the end effector corresponding to the manipulation of the item and an operation of the robotic arm corresponding to the movement of the end effector;
receive the multi-dimensional time-series data of the set of sensors;
input the multi-dimensional time-series data to a classifier of the computer system, the classifier trained for state-based event classifications;
receive, from the classifier and based at least in part on processing the multi-dimensional time-series data across a plurality of states of the robotic arm and on a plurality of transition likelihoods, an output that classifies the inventory action in an event classification from the state-based event classification, wherein each of the plurality of transition likelihoods corresponds to a likelihood of a transition between two of the plurality of states, the robotic arm is configured to perform a different manipulation of the item specific to each one of the plurality of states; and
store the event classification of the inventory action.
|