Abstract: Disclosed herein are systems, methods, and devices for using adaptive learning to identify objects. An object-identifying device performs a first object identification based on one or more features of a first modality of an object retrieved from an image frame including the object and a first database including first modality identification features. A second object identification is performed based on one or more features of a second modality of the object retrieved from the image frame and a second database including second modality identification features. The second database is updated by adaptively learning a new second modality identification feature according to a first identification result of the first object identification. The second object identification is trained with the updated second database and determines a final identification result by integrating a first identification result of the first object identification and a second identification result of the second object identification.

Abstract: A robot including a resource sharing circuit, configured to operate according to a resource utilization limit, and comprising a processor, configured to receive resource data representing a resource utilization of the robot; if the resource utilization of the robot is within a predetermined range, operate according to a first operational mode, wherein an upper limit of the predetermined range is defined by a tolerance relative to the resource utilization limit; if the resource utilization of the robot is outside of the predetermined range, operate according to a second operational mode; wherein the first operational mode includes controlling a communication circuit to send a first wireless signal representing an availability to accept a task; and wherein the second operational mode includes controlling the communication circuit to send a second wireless signal representing an availability to allocate a task to an external device for remote processing.

Abstract: Described herein is a technique that provides a first robot associated with a first trust domain to identify and authenticate a second robot associated with a second trust domain. The identification and authentication technique described herein is based on a robot processing a variety of input signals obtained via sensors, in order to generate robot identification information that can be compared with known or trusted information. Advantageously, the identification technique occurs locally, for example, at the robot and at the location of the robot-to-robot interaction, eliminating any requirement for communicating with an authoritative remote or cloud-based service, at the time and place of the robot-to-robot interaction.

Abstract: Disclosed herein are systems, devices, and methods of a robot safety system for improving the safety of human-robot collaborations. The robot safety system may include a processor that receives a collaboration policy for task collaboration between a robot and a collaborator. The task collaboration may include a movement plan of the robot to move a manipulator to a collaboration position. The robot safety system may also determine a new collaboration position based on the collaboration policy. The robot safety system may also update the collaboration position of the movement plan to the new collaboration position.

Abstract: Systems, apparatuses and methods may provide for technology that groups sensor data into a plurality of clusters based on feature similarity, conducts an artificial intelligence (AI) analysis of the plurality of clusters, and detects a data defect based on the AI analysis.

Abstract: A robot may include a processor, configured to operate according to a first operational mode or a second operational mode, wherein the first operational mode comprises performing an environmental perception operation using first sensor data, wherein the first sensor data comprise no camera data, and wherein the second operational mode comprises performing the environmental perception operation using second sensor data, wherein the second sensor data comprise camera data; while operating according to the first operational mode, determine a confidence factor representing a confidence of the environmental perception operation; determine whether or not the confidence factor is outside of a range; and if the confidence factor is outside of the range, operate according the second operational mode.

Abstract: Disclosed herein are systems and methods for controlling a telepresence robot, sometimes referred to as a receiver. The systems and methods may include obtaining environmental data associated with the receiver and/or an operator of the telepresence robot, sometimes referred to as a sender. A model defining a human intent may be received and an intent of a human proximate the receiver and or the sender may be determined using the model. A first signal may be transmitted to the receiver. The first signal may be operative to cause the receiver to alter a first behavior based on the intent of the human and/or the sender.

Abstract: Disclosed herein are embodiments of systems and methods for diversified imitation learning for automated machines. In an embodiment, a process-profiling system obtains sensor data captured by a plurality of sensors that are arranged to observe one or more human subjects performing one or more processes to accomplish one or more tasks. The process-profiling system clusters the sensor data based on a set of one or more process-performance criteria. The process-profiling system also performs, based on the clustered sensor data, one or both of generating and updating one or more process profiles in a plurality of process profiles. The process-profiling system selects, for one or more corresponding automated machines, one or more process profiles from among the plurality of process profiles, and the process-profiling system configures the one or more corresponding automated machines to operate according to the selected one or more process profiles.

Abstract: In one embodiment, an apparatus comprises a processor to: identify a workload comprising a plurality of tasks; generate a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks; identify a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with a plurality of processing devices; identify a privacy policy associated with the workload; identify privacy level information associated with the plurality of processing devices; identify a privacy constraint based on the privacy policy and the privacy level information; and determine a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph.

Abstract: Localization data is accessed, which is based on information collected by one or more sensor devices deployed within an environment. The localization data identifies presence of a person within a particular location within the environment. A feature vector is access, which includes values to describe a plurality of different physical characteristics of the person as measured by a set of one or more sensor devices deployed within the environment when the person was present within the particular location. An emotion of the person is determined from the feature vector data, and the emotion is associated with the particular location within the environment. An emotion heat map is generated of the environment to map emotional responses determined within the environment to specific locations within the environment.

Abstract: Techniques for monitoring the health of a device comprising a motor or generator are described. An apparatus to monitor the health of the device includes a plurality of sensors to detect root-mean-square (RMS) current values for a plurality of stator windings of the device. The apparatus also includes a processor to receive the RMS current values from the plurality of sensors and compute a fault indicator based on the RMS current values. The fault indicator represents a level of current imbalance between the plurality of stator windings due to winding faults.

Abstract: A system includes a self-learning module for creating a self-learned planogram based on images of shelving units at a location and shelving unit tracking. The self-learned planogram includes shelving unit locations for the shelving units. The system also includes a training module for training the merchandise tracking model based on merchandise-shelving unit clustering. The merchandise-shelving unit clustering is based on the self-learned planogram and sensor readings received from sensors at the location. The sensor readings are associated with items at the location. The system further includes a tracking module for tracking and storing locations of the items based on the sensor readings and the merchandise tracking model. The system also includes a planogram compliance module for determining planogram compliance based on comparing the self-learned planogram to the item locations.

Abstract: Systems and methods for double real-time bidding at a brick and mortar store are disclosed. The systems may include a shopper tracking system for tracking a plurality of shoppers, an electronic sign for displaying messages offering products for auction, and a real-time bidding system for receiving bids to sell an offered product from a plurality of sellers and bids to purchase the product from the plurality of shoppers. Shopping behavior data may be collected by the shopper tracking system from the plurality of shoppers, and used by the real-time bidding system to select the messages offering products.

Abstract: Monitoring and analyzing people interacting with an environment, e.g., looking around, touching, talking, exclaiming, moving items around, staring at something, walking, as well as producing identifiable physiological and/or movement that may be analyzed to determine a person's interest and/or emotional response to the environment. In a shopping context, analysis may determine customer interest in a specific product. Interest/emotional response may be translated into a quantified review and the reviews may be pushed peer-to-peer to others in the environment, used to update signage, propagated to social media, etc. Reviews may be determined in real time and if related to an unusual or negative review, may trigger one or more various responses, e.g., automatic assistance to clear up a problem, perform other error handling, notify a vendor of a problem, trigger an update such as to lower prices, etc.

Abstract: Localization data is accessed, which is based on information collected by one or more sensor devices deployed within an environment. The localization data identifies presence of a person within a particular location within the environment. A feature vector is access, which includes values to describe a plurality of different physical characteristics of the person as measured by a set of one or more sensor devices deployed within the environment when the person was present within the particular location. An emotion of the person is determined from the feature vector data, and the emotion is associated with the particular location within the environment. An emotion heat map is generated of the environment to map emotional responses determined within the environment to specific locations within the environment.

Abstract: In one embodiment, an apparatus comprises a top-view sensing device to capture sensor data associated with an environment below the top-view sensing device. The apparatus further comprises a processor to: obtain the sensor data captured by the top-view sensing device; generate, based on the sensor data, a visual representation of the environment below the top-view sensing device; determine that the visual representation comprises a representation of a person; identify one or more features associated with the representation of the person; and identify demographic information associated with the person based on the one or more features.

Abstract: Techniques for monitoring the health of a device comprising a motor or generator are described. An apparatus to monitor the health of the device includes a plurality of sensors to detect root-mean-square (RMS) current values for a plurality of stator windings of the device. The apparatus also includes a processor to receive the RMS current values from the plurality of sensors and compute a fault indicator based on the RMS current values. The fault indicator represents a level of current imbalance between the plurality of stator windings due to winding faults.