Abstract: Systems, methods, and computer-readable media are disclosed for a systems and methods for improved smart infrastructure data transfer. An example method may involve determining, at a first time, that a first number of smart infrastructure devices included within a smart infrastructure system are accessible over a network. The example method may also involve creating a first software image including a software update for the first number of smart infrastructure devices. The example method may also involve sending, over the network, the first software image to a device of the smart infrastructure system. The example method may also involve determining, at a second time, that a second number of smart infrastructure devices included within the smart infrastructure system are accessible over the network. The example method may also involve determining that the first number of devices is different than the second number of devices.

Abstract: A system of verifying execution sequence integrity of an execution flow includes a monitoring system in communication with one or more sensors of a system being monitored, where the monitoring system includes one or more electronic devices, and a computer-readable storage medium having one or more programming instructions. When executed, the one or more programming instructions cause at least one of the electronic devices to receive from the sensors, a parameter value for each of one or more parameters that pertain to an operational state of the system, combine the received parameters to generate a combination value, apply a hashing algorithm to the combination value to generate a temporary hash value, search a data store for a result code associated with the temporary hash value, and in response to the result code associated with the temporary hash value indicating that the temporary hash value is incorrect, generate a fault notification.

Abstract: A method of verifying execution sequence integrity of an execution flow includes receiving, by a local monitor of an automated device monitoring system from one or more sensors of an automated device, a unique identifier for each function in a subset of an execution flow for which the local monitor is responsible for monitoring. The method includes combining the received unique identifiers to generate a combination value, applying a hashing algorithm to the combination value to generate a temporary hash value, retrieving, from a data store, a true hash value, determining whether the temporary hash value matches the true hash value, and in response to the temporary hash value not matching the true hash value, generating a fault notification. The true hash value represents a result of applying the hashing algorithm to a combination of actual unique identifiers associated with each function in the subset.

Abstract: A method and a system for forecasting trajectories in an autonomous vehicle using recurrent neural networks. The method includes receiving a first set of data that comprises time series information corresponding to states of a plurality of objects, analyzing the first set of data to determine a plurality of object trajectory sequences corresponding to the plurality of objects, and using one or more of the plurality of object trajectory sequences as input to train a prediction model for predicting future trajectories of the plurality of objects. The predication model can be trained by defining a first prediction horizon, training the prediction model over the first prediction horizon to generate a semi-trained prediction model, defining a second prediction horizon that is longer than the first prediction horizon, and training the semi-trained prediction model to generate a trained prediction model.

Abstract: A system detects multiple instances of an object in a digital image by receiving a two-dimensional (2D) image that includes a plurality of instances of an object in an environment. For example, the system may receive the 2D image from a camera or other sensing modality of an autonomous vehicle (AV). The system uses a first object detection network to generate a plurality of predicted object instances in the image. The system then receives a data set that comprises depth information corresponding to the plurality of instances of the object in the environment. The data set may be received, for example, from a stereo camera of an AV, and the depth information may be in the form of a disparity map. The system may use the depth information to identify an individual instance from the plurality of predicted object instances in the image.

Abstract: An apparatus and method for joint calibration of vision and radar sensors for an autonomous device is disclosed. The apparatus may include a spherical portion and a cutout portion. The cutout portion may be formed within the spherical portion and have three equal surfaces. Additionally, the apparatus may include a trihedral reflector positioned within the cutout portion.

Abstract: Methods and systems for thermal management of sensors of a vehicle. The system includes a sensor that is attached to a vehicle, a temperature monitor configured to collect temperature data corresponding to the sensor, a processor, and a sensor cleaning subsystem. The sensor cleaning subsystem includes a cleaning fluid reservoir and a port that is positioned to direct cleaning fluid from the reservoir to the sensor. The processor is configured to: receive temperature data corresponding to the sensor from the temperature monitor, determine whether a current temperature of the sensor is less than a threshold temperature, and in response to determining that the current temperature is less than the first threshold temperature, cause the sensor cleaning subsystem to: heat the cleaning fluid, and initiate a cleaning cycle and direct the heated cleaning fluid to the sensor to heat the sensor.

Abstract: Systems and methods for operating a vehicle. The methods comprise: generating, by a computing device, a vehicle trajectory for the vehicle that is in motion; detecting an object within a given distance from the vehicle; generating, by the computing device, at least one possible object trajectory for the object which was detected; performing, by the computing device, a collision check to determine that there remains time to react safely to worst-case behavior by the object (the collision check being based on the vehicle trajectory and at least one possible object trajectory); and performing operations, by the computing device, to selectively cause the vehicle to perform an emergency maneuver based on results of the collision check.

Abstract: An autonomous mode controller of an autonomous vehicle according to various rider-selectable comfort configurations. When the controller identifies a comfort configuration, it will access a configuration data set that corresponds to the selected comfort configuration. The controller also will receive sensor data from one or more autonomous driving sensors. In response to the received sensor data, the controller will generate an instruction for operation of a steering, braking, powertrain or other subsystem for operation of the autonomous vehicle. The instruction while includes values that correspond to the sensed data and to one or more parameters of the configuration data set. The subsystem will then cause the autonomous vehicle to move according to the instruction.

Abstract: A method of scheduling a plurality of tasks in an autonomous vehicle system (AVS) includes, by a processor, prior to runtime of an autonomous vehicle, identifying a plurality of tasks to be implemented by the AVS of the autonomous vehicle, for each of the tasks, identifying at least one fixed parameter and at least one variable, and developing a schedule for each of the tasks. The schedule includes an event loop that minimizes an overall time for execution of the tasks. The method includes compiling the schedule into an execution plan, and saving the execution plan to a memory of the autonomous vehicle. During runtime of the autonomous vehicle, the processor receives data corresponding to the variables of the tasks, and uses the variables to implement the execution plan on the autonomous vehicle.

Abstract: Methods and systems are disclosed for correlating synthetic LiDAR data to a real-world domain for use in training an autonomous vehicle in how to operate in an environment. To do this, the system will obtain a data set of synthetic LiDAR data, transfer the synthetic LiDAR data to a two-dimensional representation, use the two-dimensional representation to train a model of a real-world environment, and use the trained model of the real-world environment to train an autonomous vehicle.

Abstract: Disclosed are improved LiDAR systems and methods that achieve an improved signal-to-noise by interrogating a sub-region of a scene with an optical signal. An instantaneous field-of-view (FOV) of each detector pixel is narrowed along a first direction to reduce detection of solar-generated photons. Instantaneous FOVs of the pixels are compressed along the first direction to provide a composite FOV that is narrower than a total FOV. To sample the total FOV of a scene, the optical signal and composite FOV of the receiver are scanned across the scene along the first direction.

Abstract: Methods and systems for thermal management of sensors of a vehicle. The system includes a sensor that is attached to a vehicle, a temperature monitor configured to collect temperature data corresponding to the sensor, a processor, and a sensor cleaning subsystem. The sensor cleaning subsystem includes a cleaning fluid reservoir and a port that is positioned to direct cleaning fluid from the reservoir to the sensor. The processor is configured to: receive temperature data corresponding to the sensor from the temperature monitor, determine that a current temperature of the sensor is greater than a first threshold temperature, and cause the sensor cleaning subsystem to initiate a cleaning cycle and direct the cleaning fluid to the sensor to cool down the sensor in response to determining that the current temperature of the sensor is greater than the first threshold temperature.

Abstract: A system uses video of a vehicle or other object to detect and classify an active turn sign on the object. The system generates an image stack by scaling and shifting a set of digital image frames from the video to a fixed scale, yielding a sequence of images over a time period. The system processes the image stack with a classifier to determine a pose of the object, as well as the state and class of each visible turn signals on the object. When the system determines that a turn signal is active, the system will predict an action that the object will take based on the class of that signal.

Abstract: The present disclosure is directed toward circuits for driving one or more laser diodes with a series of current pulses, where the energy required for each current pulse is generated and stored on a pulse-by-pulse basis. Laser-driver circuits in accordance with the present disclosure include a charge-storage inductor that is electrically coupled with a power supply and a charge-storage capacitor that is electrically coupled with a laser-diode string. The electrical coupling between the inductor and capacitor is controlled by one or more switches having on- and off-states that determine whether the inductor is charged by the power supply, charges the capacitor, or whether the charged capacitor generates a current pulse in the laser-diode string. By controlling the states of the switches, the energy provided to the laser-diode string can be controlled on a pulse-by-pulse basis.

Abstract: A method of estimating a kinetic state of a rigid body is disclosed. A LiDAR sensor receives multiple point clouds. A processor determined a rigid body transform for a pair of the received point clouds and obtains a sample space of previous rigid body transforms. The processor then identifies a proposal distribution that represents the probability that any rigid body transform in the sample space best aligns a subset of points of a first point cloud of the pair to a subset of points in a second point cloud of the pair. The processor then uses importance sampling to identify a probability distribution for the rigid body transform that best aligns the subsets of points, It then uses the identified probability distribution and a history of probability distributions to estimate a kinetic state of the rigid body.

Abstract: A monolithic series-connected laser-diode array is presented, where the array is formed on a non-conductive substrate that includes a plurality of discrete electrically conductive regions. Each laser diode of the array is disposed on a different conductive region such that the laser cavity of each laser diode is optically isolated from its respective conductive region, thereby avoiding optical loss in the laser cavity due to interaction with the highly doped conductive material. Each conductive region is configured to extend past the lateral extent of its respective laser-diode structure. Electrical connection between adjacent laser diodes of the array is made by forming a conductive trace that extends from the top contact of one of the laser diodes to the conductive region on which the other laser diode is disposed.

Abstract: An assembly includes a housing arranged to house a sensor. The assembly includes a ring-shaped first shield supported by the housing. The assembly includes a second shield supported by the housing above the first shield and defining an outer perimeter greater than an inner perimeter, and less than an outer perimeter, of the first shield.

Abstract: Disclosed are improved LiDAR systems methods employing late-lock Geiger mode detection. In sharp contrast to the prior art, a late-lock Geiger mode detection system and/or method utilizes a pulsed laser and asynchronous avalanche photodiodes with a holdoff time between photodiode arm pulses that are substantially equal to—but slightly less than—the laser pulse period. Preferably, such difference between the holdoff time and the pulse period is <10 nsec.

Abstract: A method for developing a map of objects in a region surrounding a location is disclosed. The method includes interrogating the region along a detection axis with a series of optical pulses and detecting reflections of the optical pulses that originate at objects located along the detection axis. A multi-dimensional map of the region is developed by scanning the detection axis about the location in at least one dimension. The reflections are detected via a single-photon detector that is armed using a sub-gating scheme such that the single-photon detector selectively detects photons of reflections that originate only within each of a plurality of zones that collectively define the detection field. In some embodiments, the optical pulses have a wavelength within the range of 1350 nm to 1390 nm, which is a spectral range having a relatively high eye-safety threshold and a relatively low solar background.