osi_logicaldetectiondata.proto

syntax = "proto2";

option optimize_for = SPEED;

import "osi_version.proto";
import "osi_common.proto";

package osi3;

//
// \brief Processed data from one or multiple sensors as a list of logical detections.
// Logical detections are derived from sensor detections in a logical model through processing steps like fusion filtering, tracking etc.
//
// All information is given with respect to the reference frame of the logical/virtual sensor
// \c SensorView::mounting_position (e.g. center of rear axle of the ego car) in Cartesian coordinates.
//
message LogicalDetectionData
{
    // The interface version used by the sender (i.e. the simulation
    // environment).
    //
    optional InterfaceVersion version = 1;

    // Header attributes of fused detections from multiple sensors and sensor types.
    //
    optional LogicalDetectionDataHeader header = 2;

    // logical detections consisting of transformed (and potentially fused) detections from one or multiple sensors and sensor types.
    // Logical detections are given with respect to the reference frame of the logical/virtual sensor
    // \c SensorView::mounting_position (e.g. center of rear axle of the ego car)
    //
    repeated LogicalDetection logical_detection = 3;
}

//
// \brief The header attributes of each sensor's logical detections list.
//
message LogicalDetectionDataHeader
{
    // Time stamp at which the transformation and optional fusion was finished in the global synchronized time.
    //
    // \note See \c SensorData::timestamp and
    // \c SensorData::last_measurement_time for detailed discussions on the
    // semantics of time-related fields.
    //
    optional Timestamp logical_detection_time = 1;

    // Data Qualifier expresses to what extent the content of this event can be
    // relied on.
    //
    optional DataQualifier data_qualifier = 2;

    // The current number of valid detections in the logical detections list.
    //
    // \note This value has to be set if the list contains invalid logical detections.
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // \endrules
    //
    optional uint32 number_of_valid_logical_detections = 3;

    // The ID(s) of the sensor(s) that produced the detections for transformation
    // and - in case of multiple sensors - fusion into logical detections.
    //
    repeated Identifier sensor_id = 4;

    // Data qualifier communicates the overall availability of the
    // interface.
    //
    enum DataQualifier
    {
        // Unknown (must not be used in ground truth).
        //
        DATA_QUALIFIER_UNKNOWN = 0;

        // Other (unspecified but known).
        //
        DATA_QUALIFIER_OTHER = 1;

        // Data is available.
        //
        DATA_QUALIFIER_AVAILABLE = 2;

        // Reduced data is available.
        //
        DATA_QUALIFIER_AVAILABLE_REDUCED = 3;

        // Data is not available.
        //
        DATA_QUALIFIER_NOT_AVAILABLE = 4;

        // Sensor is blind.
        //
        DATA_QUALIFIER_BLINDNESS = 5;

        // Sensor temporary available.
        //
        DATA_QUALIFIER_TEMPORARY_AVAILABLE = 6;

        // Sensor invalid.
        //
        DATA_QUALIFIER_INVALID = 7;
    }
}

//
// \brief A logical detection that could be based on multiple sensors and sensor types.
//
message LogicalDetection
{
    // Existence probability of the logical detection
    //
    // \note Use as confidence measure where a low value means less confidence
    // and a high value indicates strong confidence.
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // is_less_than_or_equal_to: 1
    // \endrules
    //
    optional double existence_probability = 1;

    // ID of the detected object this logical detection is associated to.
    //
    // \note ID = MAX(uint64) indicates no reference to an object.
    //
    // \rules
    // refers_to: DetectedObject
    // \endrules
    //
    optional Identifier object_id = 2;

    // Measured position of the logical detection given in cartesian coordinates in the
    // host vehicle coordinate system.
    //
    // Unit: m
    //
    optional Vector3d position = 3;

    // Root mean squared error of the measured position of the logical detection.
    //
    optional Vector3d position_rmse = 4;

    // Velocity of the logical detection given in cartesian coordinates in the
    // host vehicle coordinate system.
    //
    // Unit: m/s
    //
    optional Vector3d velocity = 5;

    // Root mean squared error of the logical detection's velocity.
    //
    // Unit: m/s
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // \endrules
    //
    optional Vector3d velocity_rmse = 6;

    // Intensity or equivalent value of the logical detection's echo.
    //
    // Unit: %
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // is_less_than_or_equal_to: 100
    // \endrules
    //
    optional double intensity = 7;

    // The signal to noise ratio (SNR) of the logical detection.
    //
    // Unit: dB
    //
    optional double snr = 8;

    // Describes the possibility whether more than one object may have led to
    // this logical detection.
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // is_less_than_or_equal_to: 1
    // \endrules
    //
    optional double point_target_probability = 9;

    // The ID(s) of the sensor(s) that produced the detection(s) for transformation
    // and - in case of multiple sensors - fusion into the single logical detection.
    //
    // \note One logical detection can originate from multiple sensors.
    //
    repeated Identifier sensor_id = 10;

    // Basic classification of the logical detection.
    //
    optional LogicalDetectionClassification classification = 11;

    // Echo pulse width of the logical detection's echo.
    // Several sensors output an echo pulse width instead of an intensity for each individual detection.
    // The echo pulse is measured in m and measures the extent of the object parts or atmospheric particles that produce the echo.
    // \note For more details see [1] Fig. 7 and 8.
    // \note Fig. 7 shows an example where the two echos are reflected from the edges A-B and C-D.
    // \note Fig. 8 shows how the echo pulse width is measured as the range between the rising edge and the falling edge that crosses the intensity threshold.
    //
    // Unit: m
    //
    // \rules
    // is_greater_than_or_equal_to: 0
    // \endrules
    //
    // \par Reference:
    // [1] Rosenberger, P., Holder, M.F., Cianciaruso, N. et al. (2020). <em>Sequential lidar sensor system simulation: a modular approach for simulation-based safety validation of automated driving</em> Automot. Engine Technol. 5, Fig 7, Fig 8. Retrieved May 10, 2021, from https://doi.org/10.1007/s41104-020-00066-x
    //
    optional double echo_pulse_width = 12;
}

// Definition of basic logical detection classifications.
//
enum LogicalDetectionClassification
{
    // Logical detection is unknown (must not be used in ground truth).
    //
    LOGICAL_DETECTION_CLASSIFICATION_UNKNOWN = 0;

    // Other (unspecified but known) logical detection.
    //
    LOGICAL_DETECTION_CLASSIFICATION_OTHER = 1;

    // Invalid logical detection, not to be used for object tracking, of unspecified
    // type (none of the other types applies).
    //
    LOGICAL_DETECTION_CLASSIFICATION_INVALID = 2;

    // Clutter (noise, spray, rain, fog etc.).
    //
    LOGICAL_DETECTION_CLASSIFICATION_CLUTTER = 3;

    // Over-drivable (ground etc.).
    //
    LOGICAL_DETECTION_CLASSIFICATION_OVERDRIVABLE = 4;

    // Under-drivable (sign gantry etc.).
    //
    LOGICAL_DETECTION_CLASSIFICATION_UNDERDRIVABLE = 5;
}