Abstract: An apparatus to facilitate learning reliable keypoints in situ with introspective self-supervision is disclosed. The apparatus includes one or more processors to provide a view-overlapped keyframe pair from a pose graph that is generated by a visual simultaneous localization and mapping (VSLAM) process executed by the one or more processors: determine a keypoint match from the view-overlapped key frame pair based on a keypoint detection and matching process, the keypoint match corresponding to a keypoint: calculate an inverse reliability score based on matched pixels corresponding to the keypoint match in the view-overlapped keyframe pair: identify a supervision signal associated with the keypoint match, the supervision signal comprising a keypoint reliability score of the keypoint based on a final pose output of the VSLAM process; and train a keypoint detection neural network using the keypoint match, the inverse reliability score, and the keypoint reliability score.

Abstract: The disclosure provides techniques for map optimization for a localization and mapping system. The map optimization method includes segmenting, based on a preset segmentation condition, a trajectory tracked by the localization and mapping system to obtain a plurality of segments of the trajectory, each segment being partitioned into a head part, an interior part and a tail part; performing a global optimization process based on frames in the head and tail parts of each segment to obtain optimized mapping results for the frames in the head and tail parts of the segment; estimating optimized mapping results for frames in the interior part of each segment based on the optimized mapping results for the frames in the head and tail parts of the segment; and updating a map built by the localization and mapping system according to the optimized mapping results for the frames in each segment.

Abstract: The disclosure provides techniques for registering a camera into a map via a robot. A method for registering a camera into a map may include: collecting an observation of a calibration marker from an image captured by the camera, the calibration maker being attached to a robot capable of moving in a view of the camera; calculating, for the observation of the calibration marker, a transform T (camera, marker) between a camera frame of the camera and a marker frame of the calibration marker; obtaining, for the observation of the calibration marker, a transform T (world, robot) between a world frame of the map and a robot frame of the robot; and performing, when more than a predetermined number of observations of the calibration marker are collected during movement of the robot, a calibration process to calculate a transform T(world, camera) between the world frame and the camera frame based on the transform T (camera, marker) and the transform T(world, robot) for each observation of the calibration marker.

Abstract: A method for re-localization of the robot may include retrieving, for each of keyframes in a keyframe database of the robot, image features and a pose of the keyframe, the image features of the keyframe comprising a global descriptor and local descriptors of the keyframe (210); extracting image features of a current frame captured by the robot, the image features of the current frame comprising a global descriptor and local descriptors of the current frame (220); determining one or more rough matching frames from the keyframes based on comparison between the global descriptor of each keyframe and the global descriptor of the current frame (230); determining a final matching frame from the one or more rough matching frames based on comparison between the local descriptors of each rough matching frame and the local descriptors of the current frame (240); and calculating a pose of the current frame based on a pose of the current frame based on a pose of the final matching frame (250).

Abstract: Training data generators and methods for machine learning are disclosed. An example method to generate training data for machine learning by generating simulated training data for a target neural network, transforming, with a training data transformer, the simulated training data form transformed training data, the training data transformer trained to increase a conformance of the transformed training data and the simulated training data, and training the target neural network with the transformed training data.

Abstract: Methods and apparatus relating to FPGA (Field-Programmable Gate Array) based acceleration in robot motion planning are described. In an embodiment, logic circuitry (such as an FPGA), coupled to a processor, accelerates one or more motion planning operations for a plurality of objects. A first memory, coupled to the logic circuitry, stores data corresponding to a plurality of Oriented Bounding Boxes (OBBs). The plurality of OBBs are to provide Bounding Volume (BV) models for the plurality of objects. Other embodiments are also disclosed and claimed.

Abstract: Described herein are scene reconstruction methods and techniques for reconstructing scenes by modeling planar areas using 2.5D models and non-planar areas with 3D models. In particular, depth data for an indoor scene is received. Planar areas of the indoor scene are identified based on the depth data and modeled using a 2.5D planar model. Other areas are modeled using 3D models and the entire scene is reconstructed using both the 2.5D models and the 3D models.

Abstract: Apparatus for determining a current pose of a mobile autonomous apparatus is presented. In embodiments, an apparatus may include interface circuitry to receive detection and ranging data outputted by a Light Detection and Ranging (LIDAR) sensor that nominally sweeps and provides D degrees of detection and ranging data in continuous plurality of quanta, each covering a portion of the D degrees sweep, every time period T. The apparatus may further include pose estimation circuitry coupled to the interface circuitry to determine and provide a current pose of the mobile autonomous apparatus every fractional time period t, independent of when the LIDAR sensor actually completes each sweep. In embodiments, the apparatus may be disposed on the mobile autonomous apparatus.

Abstract: Methods, apparatus, systems, and articles of manufacture to simulate sensor data are disclosed. An example apparatus includes a noise characteristic identifier to extract a noise characteristic associated with a feature present in first sensor data obtained by a physical sensor. A feature identifier is to identify a feature present in second sensor data. The second sensor data is generated by an environment simulator simulating a virtual representation of the real sensor. A noise simulator is to synthesize noise-adjusted simulated sensor data based on the feature identified in the second sensor data and the noise characteristic associated with the feature present in the first sensor data.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to improve resource utilization for binary tree structures. An example apparatus to improve resource utilization for field programmable gate array (FPGA) resources includes a computation determiner to identify a computation capability value associated with the FPGA resources, a k-ary tree builder to build a first k-ary tree having a number of k-ary nodes equal to the computation capability value, and an FPGA memory controller to initiate collision computation by transferring the first k-ary tree to a first memory of the FPGA resources.

Abstract: A time-aware occupancy mapping using regression to unknown (“RTU”) analysis, and an apparatus to dynamically allocate occupancy probability to a cell in an environment to thereby form a time-aware occupancy map of the environment are disclosed. The apparatus includes a memory circuitry in communication with a processor circuitry, the memory circuitry configured to receive and store probability information from the processor circuitry and to store the probability value and its corresponding time stamp at a probability table.

Abstract: Methods and apparatus relating to FPGA (Field-Programmable Gate Array) based acceleration in robot motion planning are described. In an embodiment, logic circuitry (such as an FPGA), coupled to a processor, accelerates one or more motion planning operations for a plurality of objects. A first memory, coupled to the logic circuitry, stores data corresponding to a plurality of Oriented Bounding Boxes (OBBs). The plurality of OBBs are to provide Bounding Volume (BV) models for the plurality of objects. Other embodiments are also disclosed and claimed.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to improve resource utilization for binary tree structures. An example apparatus to improve resource utilization for field programmable gate array (FPGA) resources includes a computation determiner to identify a computation capability value associated with the FPGA resources, a k-ary tree builder to build a first k-ary tree having a number of k-ary nodes equal to the computation capability value, and an FPGA memory controller to initiate collision computation by transferring the first k-ary tree to a first memory of the FPGA resources.

Abstract: Apparatus for determining a current pose of a mobile autonomous apparatus is presented. In embodiments, an apparatus may include interface circuitry to receive detection and ranging data outputted by a Light Detection and Ranging (LIDAR) sensor that nominally sweeps and provides D degrees of detection and ranging data in continuous plurality of quanta, each covering a portion of the D degrees sweep, every time period T. The apparatus may further include pose estimation circuitry coupled to the interface circuitry to determine and provide a current pose of the mobile autonomous apparatus every fractional time period t, independent of when the LIDAR sensor actually completes each sweep. In embodiments, the apparatus may be disposed on the mobile autonomous apparatus.

Abstract: Methods, apparatus, systems, and articles of manufacture to simulate sensor data are disclosed. An example apparatus includes a noise characteristic identifier to extract a noise characteristic associated with a feature present in first sensor data obtained by a physical sensor. A feature identifier is to identify a feature present in second sensor data. The second sensor data is generated by an environment simulator simulating a virtual representation of the real sensor. A noise simulator is to synthesize noise-adjusted simulated sensor data based on the feature identified in the second sensor data and the noise characteristic associated with the feature present in the first sensor data.