Abstract: Techniques are disclosed to derive local driving behaviors from naturalistic data, which are then incorporated as guidance into the behavioral layer of an Automated Vehicle (AV) for adaptation to local traffic. Moreover, the local driving behaviors are implemented in the AV performance validation process. The techniques facilitate scaling of this process via automatic crowdsourced behavioral data aggregation from human-driven vehicles, as well as ADAS vehicles and autonomous vehicles, and map information. The disclosed techniques also enable the creation of a spatial-behavioral relational database that provides interfaces for efficiently querying geo-bounded local driving information, enabling customization of automated vehicle driving policies to local norms, and enabling traffic behavior analysis.

Abstract: A method for creating a road user spatio-temporal representation or threat map includes obtaining electronic map data for a spatial region and a plurality of map layers. Creating a map layer includes setting parameter(s) for a vehicle with respect to the map layer. For each subsection of the spatial region, creating the map layer includes defining a position and heading for the vehicle for each of the respective subsections and representing at least one object in the respective subsection using one or more probabilistic distributions with respect to at least velocity and position of the at least one object, and determining a collision risk value between the ego vehicle and the at least one object. The threat map is generated from the map layers with maximum acceptable collision risk values from the map layers.

Abstract: According to various embodiments, a method for operating a vehicle may include determining a vehicular area having traffic conditions or characteristics different from traffic conditions of a current or previous location of the vehicle; obtaining traffic and driving information for the determined vehicular region; changing or updating one or more of driving model parameters of a safety driving model during operation of the vehicle based on the obtained traffic and driving information; and controlling the vehicle to operate in accordance with the safety driving model using the one or more changed or updated driving model parameters. A vehicle may seamlessly update operational rules and/or handover of traffic and driving information for transitioning from one region to another.

Abstract: A safety system for a vehicle may include one or more processors configured to determine, based on a friction prediction model, one or more predictive friction coefficients between the ground and one or more tires of the ground vehicle using first ground condition data and second ground condition data. The first ground condition data represent conditions of the ground at or near the position of the ground vehicle, and the second ground condition data represent conditions of the ground in front of the ground vehicle with respect to a driving direction of the ground vehicle. The one or more processors are further configured to determine driving conditions of the ground vehicle using the determined one or more predictive friction coefficients.

Abstract: A vehicle control system, including an in-vehicle bus and a plurality of electronic control units (ECUs) coupled to the in-vehicle bus, wherein at least one ECU of the plurality of ECUs is configured to: receive, at a respective at least one ECU of the plurality of ECUs, a message in a message stream on the in-vehicle bus; evaluate the message to determine at least one of a confidence value of the security classification, a significance value of the message, or a bounds check value of the message; and determine in real-time to allow or deny the message to the vehicle control system based on at least one of the significance value of the message, the bounds check value of the message, or the confidence value of the security classification of the message, to provide a sanitized message stream to the vehicle control system.

Abstract: In an autonomous vehicle system, data received from one or more autonomous vehicles (AVs) can be aggregated to generate aggregated data. From this aggregated data, a vector-field map can be generated that includes a plurality of cells. Each of the cells can include a corresponding vector. The vector field map can be analyzed to identify one or more vectors of the plurality of cells that exceed one or more predetermined threshold values. The analysis can include a magnitude analysis and/or a frequency analysis. Based on the analysis, traffic and/or road conditions can be determined, which can provide prior knowledge about the driving behavior of other vehicles. Advantageously, aspects of the disclosure improve predictive motion models and enhance navigation algorithms.

Abstract: Technologies for thermal enhanced semantic segmentation include a computing device having a visible light camera and an infrared camera. The computing device receives a visible light image of a scene from the visible light camera and an infrared image of the scene from the infrared camera. The computing device registers the infrared image to the visible light image to generate a registered image. Registering the infrared image may include increasing resolution of the infrared image. The computing device generates a thermal boundary saliency image based on the registered infrared image. The computing device may generate the thermal boundary saliency image by applying a Gabor jet convolution to the registered infrared image. The computing device performs semantic segmentation on the visible light image, the registered infrared image, and the thermal boundary saliency image to generate a pixelwise semantic classification of the scene. Other embodiments are described and claimed.

Abstract: An Autonomous Vehicle (AV) system, including: a tracking subsystem configured to detect and track relative positioning of another vehicle that is behind or lateral to an AV configured to comply with a safety driving model, and to check a safety driving model compliance status of the other vehicle; and a risk reduction subsystem configured to plan, based on the safety driving model compliance status of the other vehicle, an AV action, wherein if the safety driving model compliance status of the other vehicle is unknown or is known to be non-compliant, the AV action is administration of a safety driving model compliance test to the other vehicle, or is a maneuver by the AV to reduce risk of collision with a leading vehicle positioned in front of the AV.

Abstract: Techniques are disclosed to realize a self-adaptive multiresolution digital plate (SAMDP) system that facilitates sharing of vehicle state information via a dynamically generated binary two-dimensional multispectral pattern. The binary 2D multispectral pattern adjusts its resolution to adapt to the visual perception process while communicating vehicle state information without relying on wireless communications (i.e. V2V, V2X, etc.). The techniques as described herein leverage the use of a multichannel spectral means that helps lower the error present in sensing and representation models used by automated driving systems.

Abstract: Technologies for management of data layers in a heterogeneous geographic information system (GIS) map are disclosed. A compute device may maintain a GIS database that includes geo-quads that represent physical locations of various scales. Data layers and layer tracks may be dynamically added to the GIS database at different scales, allowing for an extensible framework that enables a mechanism for integrating additional functionality. In the illustrative embodiment, a graph database is used to store the GIS database, allowing for a flexible structure. In some embodiments, entries in layer tracks may include binary large objects that may have properties and associated methods, allowing for application-specific functionality.

Abstract: An example includes obtaining first sensor data from a first sensor and second sensor data from a second sensor, the first sensor of a first sensor type different than a second sensor type of the second sensor; generating first encoded sensor data based on the first sensor data and second encoded sensor data based on the second sensor data; generating a contextual fused sensor data representation of the first and second sensor data based on the first and second encoded sensor data; generating first and second reconstructed sensor data based on the contextual fused sensor data representation; determining a deviation estimation based on the first and second reconstructed sensor data, the deviation estimation representative of a deviation between: (a) the first reconstructed sensor data, and (b) the first sensor data; and detecting an anomaly in the deviation estimation, the anomaly indicative of an error associated with the first sensor.

Abstract: Systems, apparatuses and methods may provide for technology that obtains a neural network output, which estimates a difference between a first measured output of a sensor and a simulated output of the sensor. The technology may also add the difference to the simulated output of the sensor. In one example, the neural network output includes mean displacement data and parametrically controllable covariance data. Additionally, the technology may subtract a point-wise difference from a second measurement output of the sensor.