Abstract: A method for training a machine learning model to derive a movement vector for a robot from image data. The method includes acquiring images from a perspective of an end-effector of the robot, forming training image data elements from the acquired images, generating augmentations of the training image data elements, training an encoder network using contrastive loss and training a neural network to reduce a loss between movement vectors output by the neural network in response to embedding outputs provided by the encoder network and respective ground truth movement vectors.

Abstract: A method for training a neural network to derive, from an image of a camera mounted on a robot, a movement vector to insert an object into an insertion. The method includes, for a plurality of positions in which the object held by the robot touches a plane in which the insertion is located controlling the robot to move to the position, taking a camera image by the camera and labelling the camera image with a movement vector between the position and the insertion in the plane and training the neural network using the labelled camera images.

Abstract: A method for controlling a robot to perform a task. The method includes acquiring, for each target of a sequence of targets comprising at least one intermediate target of the task and a final target of a task, a target image data element comprising at least one target image from a perspective of an end-effector of the robot at a respective target position of the robot and successively according to the sequence of targets, for each target in the sequence, acquiring, for the target, an origin image data element, supplying the origin image data element and the target image data to a machine learning model configured to derive a delta movement between the origin current position and the target position and controlling the robot to move according to the delta movement.

Abstract: A method for controlling a robot to perform a task. The method includes acquiring a target image data element comprising at least one target image from a perspective of an end-effector of the robot at a target position of the robot in which the robot has performed the task, acquiring an origin image data element comprising at least one origin image from the perspective of the end-effector of the robot at an origin position of the robot, supplying the origin image data element and the target image data element to a machine learning model configured to derive a delta movement between the origin current position and the target position and controlling the robot to move according to the delta movement to perform the task.

Abstract: A method for training a machine learning model to derive a movement vector for a robot from image data. The method includes acquiring images from a perspective of an end-effector of the robot, forming training image data elements from the acquired images, generating augmentations of the training image data elements, training an encoder network using contrastive loss and training a neural network to reduce a loss between movement vectors output by the neural network in response to embedding outputs provided by the encoder network and respective ground truth movement vectors.

Abstract: A method for controlling a robot to insert an object into an insertion. The method includes controlling the robot to hold the object, generating an estimate of a target position to insert the object into the insertion, controlling the robot to move to the estimated target position, taking a camera image using a camera mounted on the robot after having controlled the robot to move to the estimated target position, feeding the camera image into a neural network which is trained to derive, from camera images, movement vectors which specify movements from the positions at which the camera images are taken to insert objects into insertions and controlling the robot to move according to the movement vector derived by the neural network from the camera image.

Abstract: A method for training a neural network to derive, from an image of a camera mounted on a robot, a movement vector for the robot to insert an object into an insertion. The method includes controlling the robot to hold the object, bringing the robot into a target position in which the object is inserted in the insertion, for a plurality of positions different from the target position controlling the robot to move away from the target position to the position, taking a camera image by the camera and labelling the camera image by a movement vector to move back from the position to the target position and training the neural network using the labelled camera images.

Abstract: A method for training a neural network to derive, from an image of a camera mounted on a robot, a movement vector to insert an object into an insertion. The method includes, for a plurality of positions in which the object held by the robot touches a plane in which the insertion is located controlling the robot to move to the position, taking a camera image by the camera and labelling the camera image with a movement vector between the position and the insertion in the plane and training the neural network using the labelled camera images.

Abstract: A method for training a neural network to derive, from a force and a moment exerted on an object when pressed on a plane in which an insertion for inserting the object is located, a movement vector to insert an object into an insertion. The method includes, for a plurality of positions in which the object or the part of the object held by the robot touches a plane in which the insertion is located, controlling the robot to move to the position, controlling the robot to press the object onto the plane, measuring the force and moment experienced by the object, scaling the pair of force and moment by a number randomly chosen between zero and a predetermined positive maximum number and labelling the scaled pair by a movement vector between the position and the insertion, and training the neural network using the labelled pairs of force and moment.

Abstract: A traffic controller for a data network that includes a plurality of network nodes, a plurality of network links connecting the network nodes, and one or more edge routers, each edge router being configured to control network traffic based on permitted link capacities, and wherein one or more sources of downstream traffic data enter the network downstream of the one or more edge routers, the traffic controller including a receiver operable to periodically receive downstream transmission byte counts from at least some of the network nodes, a processor coupled with the receiver, operable to periodically update the permitted link capacities based on the network node downstream byte counts received by the receiver, and a transmitter coupled with the processor operable to periodically transmit the thus-updated permitted link capacities to the one or more edge routers for their use in controlling the network traffic.

Abstract: Methods for dynamic bandwidth allocation among optical networks units for upstream transmission, performed by an optical line terminal of a passive optical network, which compensate for unused guaranteed bandwidth, which compensate for inability to use allocated bandwidth due to lack of data frame fragmentation, and which provide fair bandwidth allocation between mixed high transmission rate and low transmission rate optical network units.

Abstract: A power manager for a passive optical network, including a network statistics collector, for collecting data regarding traffic in a passive optical network (PON) including a plurality of optical network units (ONUs) on the downstream side of the PON, wherein each ONU can be in at least a sleep state and an active state, and wherein the PON transmits data in packets of data frames, a buffer for storing downstream data frames for each ONU while the ONU is in the sleep state, an activity detector for processing the data collected by the network statistics collector to generate indicators of activity levels for each ONU, and a protocol manager including a plurality of state machines for the respective plurality of ONUs, wherein each state machine governs state transition of its respective ONU to the sleep state when the activity detector indicates a low activity level for the ONU, and to the active state when the activity detector indicates a high level of activity for the ONU.

Abstract: Methods for dynamic bandwidth allocation among optical networks units for upstream transmission, performed by an optical line terminal of a passive optical network, which compensate for unused guaranteed bandwidth, which compensate for inability to use allocated bandwidth due to lack of data frame fragmentation, and which provide fair bandwidth allocation between mixed high transmission rate and low transmission rate optical network units.

Abstract: An OLT allocates a bandwidth budget and assigns upstream transmission order by receiving upstream transmission requests from a plurality of ONUs. Each ONU's request includes a requested guaranteed bandwidth and a requested best effort bandwidth. Each ONU has respective first and second attribute values. One attribute is given allocation priority over the other attribute. One attribute is given scheduling priority over the other attribute. Within each attribute, an allocation rank and a transmission rank is assigned to the possible attribute values. The bandwidth budget is allocated in accordance with the allocation priority and ranks. The upstream transmissions are scheduled in accordance with the scheduling priority and ranks.

Abstract: A power manager for a passive optical network, including a network statistics collector, for collecting data regarding traffic in a passive optical network (PON) including a plurality of optical network units (ONUs) on the downstream side of the PON, wherein each ONU can be in at least a sleep state and an active state, and wherein the PON transmits data in packets of data frames, a buffer for storing downstream data frames for each ONU while the ONU is in the sleep state, an activity detector for processing the data collected by the network statistics collector to generate indicators of activity levels for each ONU, and a protocol manager including a plurality of state machines for the respective plurality of ONUs, wherein each state machine governs state transition of its respective ONU to the sleep state when the activity detector indicates a low activity level for the ONU, and to the active state when the activity detector indicates a high level of activity for the ONU.

Abstract: An OLT allocates a bandwidth budget and assigns upstream transmission order by receiving upstream transmission requests from a plurality of ONUs. Each ONU's request includes a requested guaranteed bandwidth and a requested best effort bandwidth. Each ONU has respective first and second attribute values. One attribute is given allocation priority over the other attribute. One attribute is given scheduling priority over the other attribute. Within each attribute, an allocation rank and a transmission rank is assigned to the possible attribute values. The bandwidth budget is allocated in accordance with the allocation priority and ranks. The upstream transmissions are scheduled in accordance with the scheduling priority and ranks.

Abstract: The present invention discloses devices and methods for identifying, analyzing, and repairing network problems. The present invention can be implemented various types of networks including a packet-switched network and an Ethernet passive optical network (EPON), caused by a variety of reasons, culminating in undesirable packet discard. A method of discard-sniffing (DS) is disclosed for monitoring discarded network traffic. A discard reason register stores discard decisions and indications for discard-designated frames, allowing a network administrator to analyze causes for frames being discarded. The discard-designated frames can be routed to alternate destinations based on their designation. Optionally, a configuration register is available to disable DS capability.