# Message API Overview

## Message API Overview

Section Message Description
actor_group ActorGroupProto A message to request creation and destruction of actors in simulator
actuator_group ActuatorGroupProto A message used to send actuator commands to hardware (or simulated hardware)
alice LogMessageProto Header for logging messages
MessageChannelIndexProto An index containing all message channels in a log
MessageHeaderProto A message header which is sent together with the actual message. It allows the remote to forward the message to the correct receiver and it also provides timing information based on the transmitter clock.
PoseTreeEdgeProto Information about the pose releation between two coordinate frames.
UdpPackageHeaderProto A small header for every package we send over a UDP socket. It allows the remote to reassemble the message
audio_data AudioDataProto Message containing input to the audio playback module
audio_file_playback AudioFilePlaybackProto Notifies the audio file loader module the file to play
DepthCameraProto Depth information for an image
DisparityCameraProto Disparity information for an image
DistortionProto Distortion parameters for a camera image
PinholeProto Pinhole camera model parameters
SegmentationCameraProto Pixel-wise class label and instance segmentations for an image
chat_message ChatMessageProto A chat message
collision CollisionProto A collision event information
detections ConfusionMatrixProto A message to contain the evaluation results of an object detection algorithm. Contains the confusion matrices for the classes in a dataset accumulated over a certain number of images in the dataset.
Detections2Proto A message containing detections made by sensor(s), in the 2D image space Each detection has a bounding box and/or 2D pose, label, and confidence
Detections3Proto A message containing detections made by sensor(s), in the 3D space Each detection has a 3D pose, label, and confidence
PredictionProto A prediction gives a class name with a confidence level from 0 to 1
Skeleton2ListProto Defines a list of Skeleton2Proto messages.
Skeleton2Proto Defines a skeleton with 2D joint locations. Used for example to describe a human skeleton detected on an image.
differential_base DifferentialState Describes how an entity is moving through space in a local diverging coordinate frame.
DifferentialTrajectoryPlanProto A trajectory plan for a differential base consisting of a sequence of states
Goal2FeedbackProto A status update which is sent continously as a reply to receiving a goal message. This can be used to keep track if the robot has arrived at the destination.
Goal2Proto Describe a goal in a given referential frame
GoalWaypointProto A desired target waypoint
Odometry2Proto Describes how an entity is moving through space in a local diverging coordinate frame.
Plan2Proto Describes a plan in a given referential frame.
dynamixel_motors DynamixelMotorsProto Commands commands for a set of dynamixel motors running for example as a dasy chain.
fiducial_list FiducialListProto A list of detected ficucials
FiducialProto A fiducial is an object placed in the field of view of an imaging system for use as a point of reference or a measure
flatscan FlatscanProto A 2D range scan which is essentially a flat version of the 3D RangeScanProto
geometry GroundPlaneProto (deprecated) A message describing the location of the ground.
PlaneProto A plane in three-dimensional space
RectangleProto A rectangle using 64-bit doubles
graph GraphProto A general graph
heatmap HeatmapProto Describes the heatmap of probabilities for different units
joystick_state JoystickStateProto Messages published from a gamepad controller.
json JsonProto A proto transporting a JSON string encoded as a string
led_strip LedStripProto A message to describe a desired state of a LED display strip
map LatticeProto The information of a lattice coordinate system: it contains the name of frame registered in the PoseTree and the size of the cells.
marker_list MarkerListProto Representation of raw marker info tracked by a motion capture system
math Matrix3dProto A matrix with dimensions 3x3 using 64-bit floating points
Pose2MeanAndCovariance A type containing mean and covariance for a Pose2
Pose2Samples A list of samples of type Pose2
Pose2dProto A pose for 2-dimensional space with rotation and translation using 64-bit doubles
Pose2fProto A pose for 2-dimensional space with rotation and translation using 32-bit floats
Pose3dProto A pose for 3-dimensional space with rotation and translation using 64-bit doubles
Pose3fProto A pose for 3-dimensional space with rotation and translation using 32-bit floats
QuaterniondProto A quaternion using 64-bit doubles for example to represent rotations
QuaternionfProto A quaternion using 32-bit floats for example to represent rotations
SO2dProto A rotation in 2-dimensional Euclidean space using 64-bit doubles
SO2fProto A rotation in 2-dimensional Euclidean space using 32-bit floats
SO3dProto A rotation in 3-dimensional Euclidean space using 64-bit doubles
SO3fProto A rotation in 3-dimensional Euclidean space using 32-bit floats
Vector2dProto A 2-dimensional vector using 64-bit doubles
Vector2fProto A 2-dimensional vector using 32-bit floats
Vector2iProto A 2-dimensional vector using 32-bit integers
Vector3dProto A 3-dimensional vector using 64-bit doubles
Vector3fProto A 3-dimensional vector using 32-bit floats
Vector3iProto A 3-dimensional vector using 32-bit integers
Vector3ubProto A 3-dimensional vector using unsigned 8-bit integers
Vector4dProto A 4-dimensional vector using 64-bit doubles
Vector4fProto A 4-dimensional vector using 32-bit floats
Vector4iProto A 4-dimensional vector using 32-bit integers
Vector5dProto A 5-dimensional vector using 64-bit doubles
Vector5fProto A 5-dimensional vector using 32-bit floats
Vector5iProto A 5-dimensional vector using 32-bit integers
obstacles ObstaclesProto A message containing all obstacles around the robot
ping PingProto A generic message to ping another node with a private message.
point_cloud PointCloudProto A message containing information about a set of points , also called “point cloud”. Points can have various attributes. This message provides the position of the point in 3D space, a normal, a color and an intensity. Point attributes are each stored as NxM tensors. N is the number points which has to be identical for all attributes or 0. M is the dimension of the corresponding attribute.
pose_tree PoseTreeProto Pose tree/kinematic tree: hierarchicial kinematic representation
pwm_channel_set PwmChannelSetDutyCycleProto Set a duty cycle for a given PWM channel
PwmChannelSetPulseLengthProto Set a pulse length for a given PWM channel
range_scan RangeScanProto A message about laser range scans published for example from a LIDAR sensor The message does not have to capture a full 360 degree scan, as most sensors publish partial slices at high frequency instead of waiting to complete a full scan.
rigid_body_3_group RigidBody3GroupProto A list of rigid bodies in a certain coordinate frame
RigidBody3Proto A rigid body in 3D
robot_state RobotStateProto The robot state at that particular instance of time. The proto stores the current position and angle, as well as the difference between the previous values of the states (position and angle). Also contains the current linear and angular speed, obtained from odometry input
segmentation_prediction SegmentationPredictionProto Segmentation prediction proto stores the probability distribution over classes for each pixel in a three dimensional tensor. The probability will add up to one for each pixel.
state StateProto A message used to transport states of dynamic systems. This is used closely together with the state gem to be found in //engine/gems/state. For example you can define the robot state using the compile-time types defined in that gem, and then serialize that state into a message. The receiver of that message can deserialize the state back into the compile-time robot state type.
superpixels SuperpixelLabelsProto Labels for superpixels. Assigns a label to every superpixel in a related SuperpixelsProto
SuperpixelsProto A superpixel oversegmentation for an image. This is useful for various image segmentation methods or perception algorithms in general. # Each pixel in the image is assigned to a superpixel. For each superpixel information like the pixel coordinate or the color are reported. Optionally also 3D data like point and normal are reported.
trajectory Vector3TrajectoryListProto A list of trajectories with three dimensions. For example, one might consider the multiple trajectories for given bodies.
Vector3TrajectoryProto A trajectory with three dimensions. The trajectory is composed of a time series of positions in reference to a given frame. For example, one might consider the trajectory of a moving target. Note: The timestamps may be omitted, in which case the trajectory is equivalent to a path.
uuid UuidProto A unique identifier following the UUID specifications
voice_command_detection VoiceCommandDetectionProto Message containing detected command id and list of timestamps of contributing keywords

## actor_group

### ActorGroupProto

Copy
Copied!

# A message to request creation and destruction of actors in simulator
struct ActorGroupProto {
# A request to spawn a new actor
struct SpawnRequest {
# The name given to the spawned actor. This name does not need to be unique. If multiple spawned
# actors have the same name, the subsequent requests apply to all of them.
name @0: Text;
# The object template the spawned actor is instantiated from (e.g. a prefab name in Unity)
prefab @1: Text;
# The pose to set the spawned actor (in the simulator's reference frame)
pose @2: Pose3dProto;
}
# List of spawn requests
spawnRequests @0: List(SpawnRequest);
# List of names of actors to destroy.
destroyRequests @1: List(Text);
}


## actuator_group

### ActuatorGroupProto

Copy
Copied!

# A message used to send actuator commands to hardware (or simulated hardware)
struct ActuatorGroupProto {
# The semantic meaning of the value
# Represents angular/linear values depending on the joint type
enum Semantics {
position @0;
speed @1;
acceleration @2;
force @3;
}
# A commanded property
struct Value {
semantics @0: Semantics;
# the actual control or state value
scalar @1: Float32;
}
# Just holds a list of commands per actuator
struct CommandList {
commands @0: List(Value);
}
# Desired commands per actuators
actuators @0: List(CommandList);
# (optional) names of actuators
names @1: List(Text);
}


## alice

### LogMessageProto

Copy
Copied!

# Header for logging messages
struct LogMessageProto {
acqtime @0: Int64;
pubtime @1: Int64;
contentsUuid @2: UuidProto;
}


### MessageChannelIndexProto

Copy
Copied!

# An index containing all message channels in a log
struct MessageChannelIndexProto {
struct MessageChannel {
# UUID of the app component
componentUuid @0: UuidProto;
# Tag of the message channel (as specified in ISAAC_PROTO_TX)
tag @1: Text;
# Series UUID under which messages are stored in the log
seriesUuid @2: UuidProto;
}
channels @0: List(MessageChannel);
}


Copy
Copied!

# A message header which is sent together with the actual message.
# It allows the remote to forward the message to the correct receiver and it also provides timing
# information based on the transmitter clock.
# Uniquely identifies a message across all systems
uuid @0: UuidProto;
# Uniquely identifies a proto type across all systems
proto @1: Int64;
# A channel identifier which is used to route messages
channel @2: Text;
# A (local) timestamp in nanoseconds which describes when data relevant for this message was
# aquired by the hardware.
acqtime @3: Int64;
# A (local) timestamp in nanoseconds which describes when the message was published.
pubtime @4: Int64;
messageLength @5: UInt32;
# Lengths (in bytes) of proto segments
segmentLengths @6: List(UInt32);
}


### PoseTreeEdgeProto

Copy
Copied!

# Information about the pose releation between two coordinate frames.
struct PoseTreeEdgeProto {
# Name of "left" frame, i.e. lhs for a pose lhs_T_rhs
lhs @0: Text;
# Name of "right" frame, i.e. rhs for a pose lhs_T_rhs
rhs @1: Text;
# The pose between the two frames using the orientation lhs_T_rhs
pose @2: Pose3dProto;
}


Copy
Copied!

# A small header for every package we send over a UDP socket.
# It allows the remote to reassemble the message
# Uniquely identifies a message across all systems
uuid @0: UuidProto;
# The total length of the message (header plus message payload) in bytes
messageLength @1: UInt32;
# The index of this package (assuming a fixed package size). The message header is always a
# single package which comes first.
packageIndex @2: UInt16;
}


## audio_data

### AudioDataProto

Copy
Copied!

# Message containing input to the audio playback module
struct AudioDataProto {
# Audio Sample Rate (eg. 44100, 48000, 96000, 192000)
sampleRate @0: UInt32;
# Number of channels in the Audio packet
numChannels @1: UInt8;
}


## audio_file_playback

### AudioFilePlaybackProto

Copy
Copied!

# Notifies the audio file loader module the file to play
struct AudioFilePlaybackProto {
# File index
fileIndex @0: UInt8;
}


## camera

### ColorCameraProto

Copy
Copied!

# An image published by a color (or greyscale) camera
struct ColorCameraProto {
# Actual image captured by the camera. The pixel type depends on the type of camera. Most commonly
# it could be a single 8-bit integer for a greyscale camera, or a three 8-bit integers for a
# color camera.
image @0: ImageProto;
# The choices available for color space
enum ColorSpace {
greyscale @0;
rgb @1;
bgr @2;
yuv @3;
rgba @4;
}
# Color space used by the image
colorSpace @1: ColorSpace;
# Intrinsic camera parameters
pinhole @2: PinholeProto;
distortion @3: DistortionProto;
}


### DepthCameraProto

Copy
Copied!

# Depth information for an image
struct DepthCameraProto {
# A depth value for every pixel in the image. The pixel type is a single 32-bit float.
depthImage @0: ImageProto;
# The minimum and maximum depth used
minDepth @1: Float32;
maxDepth @2: Float32;
# Intrinsic camera parameters
pinhole @3: PinholeProto;
}


### DisparityCameraProto

Copy
Copied!

# Disparity information for an image
struct DisparityCameraProto {
# A disparity value for every pixel in the image. The pixel type is a single 32-bit float.
disparityImage @0: ImageProto;
# The minimum value that a pixel could be in the disparityImage.
# This should be greater than zero so that a maximum depth can be computed.
minimumDisparity @1: Float32;
# The maximum value that a pixel could be in the disparityImage.
maximumDisparity @2: Float32;
# The distance between the two stereo cameras (in meters).
cameraBaseline @3: Float32;
# Intrinsic camera parameters
pinhole @4: PinholeProto;
}


### DistortionProto

Copy
Copied!

# Distortion parameters for a camera image
struct DistortionProto {
# The choices available for the distortion model
enum DistortionModel {
# Brown distortion model with 3 radial and 2 tangential distortion
# coefficients: (r0 r1 r2 t0 t1)
brown @0;
# Fisheye (wide-angle) distortion. 4 radial (r0, r1, r2, r3) distortion coefficients
fisheye @1;
}
# Distortion model and coefficients
model @0: DistortionModel;
coefficients @1: Vector5fProto;
}


### PinholeProto

Copy
Copied!

# Pinhole camera model parameters
struct PinholeProto {
# Resolution of the camera
rows @0: Int16;
cols @1: Int16;
# Focal length of the camera in pixel (0: row, 1: col)
focal @2: Vector2dProto;
# Optical center of the camera in pixel (0: row, 1: col)
center @3: Vector2dProto;
}


### SegmentationCameraProto

Copy
Copied!

# Pixel-wise class label and instance segmentations for an image
struct SegmentationCameraProto {
# A image which stores a class label for every pixel. The pixel type is a single unsigned 8-bit
# integer.
labelImage @0: ImageProto;
# A label used to enumerate all labels in the labels field below
struct Label {
# The integer value stored in the labelImage
index @0: UInt8;
# The name of the label
name @1: Text;
}
# List of all labels used in labelImage
labels @1: List(Label);
# A image which stores an instance index for every pixel. Different objects with the same
# label will get different instance indices. The pixel type is a single unsigned 16-bit integer.
instanceImage @2: ImageProto;
# Intrinsic camera parameters
pinhole @3: PinholeProto;
}


## chat_message

### ChatMessageProto

Copy
Copied!

# A chat message
struct ChatMessageProto {
# The user name of the sender
user @0: Text;
# A channel on which the message was posted
channel @1: Text;
# The text sent by the user
text @2: Text;
}


## collision

### CollisionProto

Copy
Copied!

# A collision event information
struct CollisionProto {
# The name of this collider
thisName @0: Text;
# The name of the other collider
otherName @1: Text;
# The pose of this collider with respect to its reference coordinate frame
thisPose @2: Pose3dProto;
# The pose of the other collider with respect to the same reference coordinate frame
otherPose @3: Pose3dProto;
# contact point of the collision in the reference coordinate frame
contactPoint @4: Vector3dProto;
# normal of the contact point in the reference coordinate frame
contactNormal @5: Vector3dProto;
# The relative linear velocity between the two colliding objects in the reference coordinate frame
velocity @6: Vector3dProto;
}


## detections

### ConfusionMatrixProto

Copy
Copied!

# A message to contain the evaluation results of an object detection algorithm. Contains the
# confusion matrices for the classes in a dataset accumulated over a certain number of images in
# the dataset.
struct ConfusionMatrixProto {
# Number of samples these metrics were accumulated over. If 1, then the confusion matrices hold
# the results for a single sample. If greater than 1, then the counts that are contained in the
# confusion matrices are summed over accumulated samples.
numSamples @0: Int64;
# List of threshold values used to compute the metrics for each class. Each threshold
# corresponds to one confusion matrix in the confusionMatrices Tensor.
# The metric on which the threshold is applied is dependent on the task. In case of object
# detection, the threshold is typically applied for the IoU (Intersection over Union) score
# between two bounding boxes.
thresholds @1: List(Float64);
# Tensor with dimensions (actual class * predicted class * IoU), where a slice of the tensor
# across a single IoU is a confusion matrix. (See https://en.wikipedia.org/wiki/Confusion_matrix)
# Each confusion matrix is calculated using the corresponding IoU threshold, which determines the
# tolerance to bounding box error. A lower IoU corresponds to greater tolerance to bbox error.
confusionMatrices @2: TensorProto;
}


### Detections2Proto

Copy
Copied!

# A message containing detections made by sensor(s), in the 2D image space
# Each detection has a bounding box and/or 2D pose, label, and confidence
struct Detections2Proto {
predictions @0: List(PredictionProto);
# List of 2D bounding boxes where we detected objects
boundingBoxes @1: List(RectangleProto);
# List of 2D poses of objects detected. Optional field.
poses @2: List(Pose2dProto);
}


### Detections3Proto

Copy
Copied!

# A message containing detections made by sensor(s), in the 3D space
# Each detection has a 3D pose, label, and confidence
struct Detections3Proto {
predictions @0: List(PredictionProto);
# List of 3D poses of objects detected relative to the sensor frame
poses @1: List(Pose3dProto);
}


### PredictionProto

Copy
Copied!

# A prediction gives a class name with a confidence level from 0 to 1
struct PredictionProto {
# Name or label for the detection
label @0: Text;
# A general value to indicate how confident we are about the detection.
# This could for example be provided by a perception algorithm like a
# neural network bounding box detector.
confidence @1: Float64;
}


### Skeleton2ListProto

Copy
Copied!

# Defines a list of Skeleton2Proto messages.
struct Skeleton2ListProto {
# List of skeleton models
skeletons @0: List(Skeleton2Proto);
}


### Skeleton2Proto

Copy
Copied!

# Defines a skeleton with 2D joint locations. Used for example to describe
# a human skeleton detected on an image.
struct Skeleton2Proto {
# A graph defining the topology of the skeleton.
graph @0: GraphProto;
# Information about a joint in the skeleton
struct Joint2Proto {
# Location of the joint, for example image pixel coordinates.
position @0: Vector2dProto;
# A label and confidence, describing the prediction of the joint
label @1: PredictionProto;
}
# Detailed information for every joint in the skeleton. The number of
# joints must be identical to the number of nodes in the graph.
joints @1: List(Joint2Proto);
# A label and confidence, describing the prediction of the skeleton
label @2: PredictionProto;
}


## differential_base

### DifferentialState

Copy
Copied!

# Describes how an entity is moving through space in a local diverging coordinate frame.
struct DifferentialState {
positionX @0: Float64;
positionY @1: Float64;
speedX @3: Float64;
speedY @4: Float64;
angularSpeed @5: Float64;
timestamp @6: Float64;
}


### DifferentialTrajectoryPlanProto

Copy
Copied!

# A trajectory plan for a differential base consisting of a sequence of states
struct DifferentialTrajectoryPlanProto {
# List of states the robot will need to be.
states @0: List(DifferentialState);
# name of the frame the plan is
planFrame @1: Text;
}


### Goal2FeedbackProto

Copy
Copied!

# A status update which is sent continously as a reply to receiving a goal message. This can be
# used to keep track if the robot has arrived at the destination.
struct Goal2FeedbackProto {
# Remaining relative pose to the goal, or identity in case hasGoal is false.
robotTGoal @0: Pose2dProto;
# Whether the robot currently has a goal
hasGoal @1: Bool;
# Whether the robot considers himself to be arrived at the target
hasArrived @2: Bool;
# Whether the robot considers himself to not move anymore
isStationary @3: Bool;
}


### Goal2Proto

Copy
Copied!

# Describe a goal in a given referential frame
struct Goal2Proto {
# The goal expressed in Pose2
goal @0: Pose2dProto;
# the tolerance radius of the goal.
tolerance @1: Float32;
# name of the frame the goal is.
goalFrame @2: Text;
# Whether or not we should stop the robot. If set to true all the other parameters will be ingored
stopRobot @3: Bool;
}


### GoalWaypointProto

Copy
Copied!

# A desired target waypoint
struct GoalWaypointProto {
# the name of the waypoint
waypoint @0: Text;
}


### Odometry2Proto

Copy
Copied!

# Describes how an entity is moving through space in a local diverging coordinate frame.
struct Odometry2Proto {
# The pose of the "robot" relative to the reference odometric frame
odomTRobot @0: Pose2dProto;
speed @1: Vector2dProto;
angularSpeed @2: Float64;
acceleration @3: Vector2dProto;
# Contains the name of the odometry frame and the robot frame.
odometryFrame @4: Text;
robotFrame @5: Text;
}


### Plan2Proto

Copy
Copied!

# Describes a plan in a given referential frame.
struct Plan2Proto {
# List of poses the robot need to go through
poses @0: List(Pose2dProto);
# name of the frame the plan is.
planFrame @1: Text;
}


## dynamixel_motors

### DynamixelMotorsProto

Copy
Copied!

# Commands commands for a set of dynamixel motors running for example as a dasy chain.
struct DynamixelMotorsProto {
# Motor protos
struct MotorProto {
# Motor ID
id @0: Int64;
# Current position
position @1: Int16;
# Tick in milliseconds since startup
tick @2: Int64;
# Is the servo moving
moving @3: Bool;
}
# A single command per motor
motors @0: List(MotorProto);
}


## fiducial_list

### FiducialListProto

Copy
Copied!

# A list of detected ficucials
struct FiducialListProto {
fiducialList @0: List(FiducialProto);
}


### FiducialProto

Copy
Copied!

# A fiducial is an object placed in the field of view
# of an imaging system for use as a point of reference or a measure
struct FiducialProto {
# A fiducial can be of type (April Tag, QRCode, Barcode or ARTag)
enum Type {
apriltag @0;
qrcode @1;
barcode @2;
artag @3;
}
# Enum to identify the type of fiducial represented by the
# proto instance
type @0: Type;
# Text field that identifies the ID of the fiducial
# For April Tag, the id is of the format <TagFamily_ID>
# Ex. If the decoded tag ID is 14 and belongs to TagFamily tag36h11,
# The id is tag36h11_14
id @1: Text;
# 3D pose of the detected tag from the camera coordinates,
# consisting of orientation (quaternion) and translation
# Camera coordinate (X - right, Y - down, Z - outward)
# Tag coordinate (X - right, Y - down, Z - opposite to tag face direction)
# Tag coordinates are centered at tag's upper-left corner
# ie. Pose has identity quaternion and zero translation, when tag is facing the camera and it's
# upper-left corner is centered at the camera center
cameraTTag @2: Pose3dProto;
# Optional list of keypoints of the detected fiducial, in image space
keypoints @3: List(Vector2dProto);
}


## flatscan

### FlatscanProto

Copy
Copied!

# A 2D range scan which is essentially a flat version of the 3D RangeScanProto
struct FlatscanProto {
# Angles (in radians) under which rays are shot
angles @0: List(Float32);
# Return distance of the ray
ranges @1: List(Float32);
# Beams with a range smaller than or equal to this distance are considered to have returned an
# invalid measurement.
invalidRangeThreshold @2: Float64;
# Beams with a range larger than or equal to this distance are considered to not have hit an
# obstacle within the maximum possible range of the sensor.
outOfRangeThreshold @3: Float64;
# Return the visibility of a given ray (the longest valid distance of a beams in this direction)
# This field is optional, however if it is set, it must have the same size as ranges and angles.
visibilities @4: List(Float32);
}


## geometry

### GroundPlaneProto

Copy
Copied!

# (deprecated) A message describing the location of the ground.
struct GroundPlaneProto {
# The ground plane.
plane @0: PlaneProto;
}


### PlaneProto

Copy
Copied!

# A plane in three-dimensional space
struct PlaneProto {
normal @0: Vector3dProto;
offset @1: Float64;
}


### RectangleProto

Copy
Copied!

# A rectangle using 64-bit doubles
struct RectangleProto {
# (x, y) values of the lower left corner of the rectangle
min @0: Vector2dProto;
# (x, y) values of the upper right corner of the rectangle
max @1: Vector2dProto;
}


## graph

### GraphProto

Copy
Copied!

# A general graph
struct GraphProto {
# The number of nodes in the graph. Indices in the list of edges will
# always be smaller than this number.
nodeCount @0: Int32;
# An edge in the graph connects two nodes.
struct EdgeProto {
source @0: Int32;
target @1: Int32;
}
# The edges of the graph.
edges @1: List(EdgeProto);
}


## heatmap

### HeatmapProto

Copy
Copied!

# Describes the heatmap of probabilities for different units
struct HeatmapProto {
# Probability for each cell of the grid
heatmap @0: ImageProto;
# Name of the frame of the map (as registered in the PoseTree)
frameName @1: Text;
# Size of a pixel in meters
gridCellSize @2: Float32;
}


## imu

### ImuProto

Copy
Copied!

# Message published by an inertial measurement unit (IMU)
struct ImuProto {
# The linear acceleration of the body frame along the primary axes
linearAccelerationX @0: Float32;
linearAccelerationY @1: Float32;
linearAccelerationZ @2: Float32;
# The angular velocity of the body frame around the primary axes
angularVelocityX @3: Float32;
angularVelocityY @4: Float32;
angularVelocityZ @5: Float32;
# Optional angles as integrated by a potential internal estimator
angleYaw @6: Float32;
anglePitch @7: Float32;
angleRoll @8: Float32;
}


## joystick_state

### JoystickStateProto

Copy
Copied!

# Messages published from a gamepad controller.
struct JoystickStateProto {
axes @0: List(Vector2dProto);
buttons @1: List(Bool);
}


## json

### JsonProto

Copy
Copied!

# A proto transporting a JSON string encoded as a string
struct JsonProto {
# The serialized JSON object
serialized @0: Text;
}


## led_strip

### LedStripProto

Copy
Copied!

# A message to describe a desired state of a LED display strip
struct LedStripProto {
# Light strip status. If false, the LED strip is powered off.
# If true, the LED strip displays the pattern and color
# defined by the following parameters of this message.
enabled @0: Bool;
# Color to display, in RGB format
color @1: Vector3ubProto;
# Number of LEDs to skip when displaying this color. The remaining
# LEDs will be unchanged.
skip @2: Int32;
# If a skip value is defined, the offset is the index of LED to start
# the skip pattern. i.e. To set every other LED
# starting from the third LED use skip = 2, offset = 3
offset @3: Int32;
}


## map

### LatticeProto

Copy
Copied!

# The information of a lattice coordinate system: it contains the name of frame registered in the
# PoseTree and the size of the cells.
struct LatticeProto {
# The name of frame as registered in the PoseTree
frame @0: Text;
# The size of a grid cell in meters
cellSize @1: Float64;
# The dimensions of the lattice grid.
dimensions @2: Vector2iProto;
}


## marker_list

### MarkerListProto

Copy
Copied!

# Representation of raw marker info tracked by a motion capture system
struct MarkerListProto {
# List of markers
markers @0: List(Marker);
# A motion captured marker
struct Marker {
# Marker name or label, if labeled in mocap system
name @0: Text;
# Name of node that this marker belongs to, if any
# If unlabeled, parent is scene root
parent @1: Text;
# Translation of marker relative to the global origin
worldTMarker @2: Vector3dProto;
# True if marker is occluded in current frame, false otherwise
occluded @3: Bool;
}
}


## math

### Matrix3dProto

Copy
Copied!

# A matrix with dimensions 3x3 using 64-bit floating points
struct Matrix3dProto {
row1 @0: Vector3dProto;
row2 @1: Vector3dProto;
row3 @2: Vector3dProto;
}


### Pose2MeanAndCovariance

Copy
Copied!

# A type containing mean and covariance for a Pose2
struct Pose2MeanAndCovariance {
# The mean of the distribution
mean @0: Pose2dProto;
# The covariance of the distribution in the reference frame of the mean.
covariance @1: Matrix3dProto;
}


### Pose2Samples

Copy
Copied!

# A list of samples of type Pose2
struct Pose2Samples {
# The state for every sample
states @0: List(Pose2dProto);
# A weight for every sample. Optional and default is equal weighting.
weights @1: List(Float64);
}


### Pose2dProto

Copy
Copied!

# A pose for 2-dimensional space with rotation and translation using 64-bit doubles
struct Pose2dProto {
rotation @0: SO2dProto;
translation @1: Vector2dProto;
}


### Pose2fProto

Copy
Copied!

# A pose for 2-dimensional space with rotation and translation using 32-bit floats
struct Pose2fProto {
rotation @0: SO2fProto;
translation @1: Vector2fProto;
}


### Pose3dProto

Copy
Copied!

# A pose for 3-dimensional space with rotation and translation using 64-bit doubles
struct Pose3dProto {
rotation @0: SO3dProto;
translation @1: Vector3dProto;
}


### Pose3fProto

Copy
Copied!

# A pose for 3-dimensional space with rotation and translation using 32-bit floats
struct Pose3fProto {
rotation @0: SO3fProto;
translation @1: Vector3fProto;
}


### QuaterniondProto

Copy
Copied!

# A quaternion using 64-bit doubles for example to represent rotations
struct QuaterniondProto {
w @0: Float64;
x @1: Float64;
y @2: Float64;
z @3: Float64;
}


### QuaternionfProto

Copy
Copied!

# A quaternion using 32-bit floats for example to represent rotations
struct QuaternionfProto {
w @0: Float32;
x @1: Float32;
y @2: Float32;
z @3: Float32;
}


### SO2dProto

Copy
Copied!

# A rotation in 2-dimensional Euclidean space using 64-bit doubles
struct SO2dProto {
# unit complex number (cos(a), sin(a)) for rotation angle a
q @0: Vector2dProto;
}


### SO2fProto

Copy
Copied!

# A rotation in 2-dimensional Euclidean space using 32-bit floats
struct SO2fProto {
# unit complex number (cos(a), sin(a)) for rotation angle a
q @0: Vector2fProto;
}


### SO3dProto

Copy
Copied!

# A rotation in 3-dimensional Euclidean space using 64-bit doubles
struct SO3dProto {
# a normalized quaternion
q @0: QuaterniondProto;
}


### SO3fProto

Copy
Copied!

# A rotation in 3-dimensional Euclidean space using 32-bit floats
struct SO3fProto {
# a normalized quaternion
q @0: QuaternionfProto;
}


### Vector2dProto

Copy
Copied!

# A 2-dimensional vector using 64-bit doubles
struct Vector2dProto {
x @0: Float64;
y @1: Float64;
}


### Vector2fProto

Copy
Copied!

# A 2-dimensional vector using 32-bit floats
struct Vector2fProto {
x @0: Float32;
y @1: Float32;
}


### Vector2iProto

Copy
Copied!

# A 2-dimensional vector using 32-bit integers
struct Vector2iProto {
x @0: Int32;
y @1: Int32;
}


### Vector3dProto

Copy
Copied!

# A 3-dimensional vector using 64-bit doubles
struct Vector3dProto {
x @0: Float64;
y @1: Float64;
z @2: Float64;
}


### Vector3fProto

Copy
Copied!

# A 3-dimensional vector using 32-bit floats
struct Vector3fProto {
x @0: Float32;
y @1: Float32;
z @2: Float32;
}


### Vector3iProto

Copy
Copied!

# A 3-dimensional vector using 32-bit integers
struct Vector3iProto {
x @0: Int32;
y @1: Int32;
z @2: Int32;
}


### Vector3ubProto

Copy
Copied!

# A 3-dimensional vector using unsigned 8-bit integers
struct Vector3ubProto {
x @0: UInt8;
y @1: UInt8;
z @2: UInt8;
}


### Vector4dProto

Copy
Copied!

# A 4-dimensional vector using 64-bit doubles
struct Vector4dProto {
x @0: Float64;
y @1: Float64;
z @2: Float64;
w @3: Float64;
}


### Vector4fProto

Copy
Copied!

# A 4-dimensional vector using 32-bit floats
struct Vector4fProto {
x @0: Float32;
y @1: Float32;
z @2: Float32;
w @3: Float32;
}


### Vector4iProto

Copy
Copied!

# A 4-dimensional vector using 32-bit integers
struct Vector4iProto {
x @0: Int32;
y @1: Int32;
z @2: Int32;
w @3: Int32;
}


### Vector5dProto

Copy
Copied!

# A 5-dimensional vector using 64-bit doubles
struct Vector5dProto {
x @0: Float64;
y @1: Float64;
z @2: Float64;
w @3: Float64;
v @4: Float64;
}


### Vector5fProto

Copy
Copied!

# A 5-dimensional vector using 32-bit floats
struct Vector5fProto {
x @0: Float32;
y @1: Float32;
z @2: Float32;
w @3: Float32;
v @4: Float32;
}


### Vector5iProto

Copy
Copied!

# A 5-dimensional vector using 32-bit integers
struct Vector5iProto {
x @0: Int32;
y @1: Int32;
z @2: Int32;
w @3: Int32;
v @4: Int32;
}


## obstacles

### ObstaclesProto

Copy
Copied!

# A message containing all obstacles around the robot
struct ObstaclesProto {
# A message containing the information about a spherical obstacle
struct SphereObstacleProto {
# The positions of the center in a given frame (see frame below)
center @0: Vector3dProto;
# The radius of the obstacle (include the minimal distance we aim to stay away from)
# The coordinate frame the center is located (must match a frame in the PoseTree)
# deprecated: use robotFrame below instead.
frame @2: Text;
# Unique identifier the can be used to track the object
uuid @3: UuidProto;
# Time this obstacle was reported (used in the PoseTree)
time @4: Float64;
}
# A distance map describes the distance to the closest obstacle for every pixel.
struct DistanceMapProto {
# The distance map as a 1d image. Each pixel contains the metric distance to the closest
# obstacle
image @0: ImageProto;
# The pose of the gridmap relative to the robot
robotTGridmap @1: Pose2dProto;
# The size of a grid cell
gridCellSize @2: Float64;
# Time this obstacle was reported (used in the PoseTree)
time @3: Float64;
}
# List of spherical obstacles (see SphereObstacleProto for details)
obstacles @0: List(SphereObstacleProto);
# List of distance maps (see DistanceMapProto for details)
distanceMaps @1: List(DistanceMapProto);
# The coordinate frame the center is located (must match a frame in the PoseTree)
robotFrame @2: Text;
}


## ping

### PingProto

Copy
Copied!

# A generic message to ping another node with a private message.
struct PingProto {
message @0: Text;
}


## point_cloud

### PointCloudProto

Copy
Copied!

# A message containing information about a set of points , also called "point cloud". Points can
# have various attributes. This message provides the position of the point in 3D space, a normal,
# a color and an intensity. Point attributes are each stored as NxM tensors. N is the number
# points which has to be identical for all attributes or 0. M is the dimension of the corresponding
# attribute.
struct PointCloudProto {
# The position of points. Most commonly XYZ stored as three 32-bit floats.
positions @0: TensorProto;
# A normal of the point. Most commonly a unit vector stored as three 32-bit floats.
normals @1: TensorProto;
# A color value for each point. Most commonly an RGB value stored as three 32-bit unit floats.
colors @2: TensorProto;
# The intensity for each point used for example by LiDAR sensors. Most commonly a single 32-bit
# float.
intensities @3: TensorProto;
}


## pose_tree

### PoseTreeProto

Copy
Copied!

# Pose tree/kinematic tree: hierarchicial kinematic representation
struct PoseTreeProto {
# Pose tree name
name @0: Text;
# List of all nodes in tree
nodes @1: List(Node);
# List of all edges between nodes in the tree, i.e. adjacency list
edges @2: List(Edge);
# Each node in the tree holds its pose relative to its immediate parent and its pose relative to the root
struct Node {
# Node name
name @0: Text;
# This node's parent index in the nodes list
parentIndex @1: Int32;
# Transformation of node relative to its parent
parentTNode @2: Pose3dProto;
# Transformation of node relative to global origin
worldTNode @3: Pose3dProto;
}
# Each edge holds a parent name and child name
# Every node knows its parent's index but this is for human readability
struct Edge {
# Parent name
parentName @0: Text;
# Child name
childName @1: Text;
}
}


## pwm_channel_set

### PwmChannelSetDutyCycleProto

Copy
Copied!

# Set a duty cycle for a given PWM channel
struct PwmChannelSetDutyCycleProto {
channel @0: Int32;
# PWM channel to set
dutyCycle @1: Float32;
# duty cycle, as a percentage (0.0 to 1.0)
disable @2: Bool;
# if set to true, duty cycle is ignored and power is set to 0
}


### PwmChannelSetPulseLengthProto

Copy
Copied!

# Set a pulse length for a given PWM channel
struct PwmChannelSetPulseLengthProto {
channel @0: Int32;
# PWM channel to set
pulseLength @1: Float32;
# length of pulse, as a percentage (0.0 to 1.0)
}


## range_scan

### RangeScanProto

Copy
Copied!

# A message about laser range scans published for example from a LIDAR sensor
# The message does not have to capture a full 360 degree scan, as most sensors publish partial
# slices at high frequency instead of waiting to complete a full scan.
struct RangeScanProto {
# Normalized distance to target (see rangeScale) of the scanned rays. Data type is 16-bit integer.
# For each angle theta there is a ray for every phi angle. So the total number of rays is
# length(theta) * length(phi). First rank is theta and second rank is phi.
ranges @0: TensorProto;
# Normalized ray return intensity (see intensityScale) of the scanned rays. Data type is 8-bit
# integer. For each angle theta there is a ray for every phi angle. So the total number of rays is
# length(theta) * length(phi). First rank is theta and second rank is phi.
# This entry is optional. Default is full intensity for all rays.
intensities @1: TensorProto;
# table of theta (horizontal) angles
theta @2: List(Float32);
# table of phi (vertical) angles
phi @3: List(Float32);
# Scale factor which can be used to convert a range value from a 16-bit integer to meters. The
# conversion formula is: range[meters] = range[normalized] / 0xFFFF * rangeScale
rangeDenormalizer @4: Float32;
# Scale factor which can be used to convert an intensity value from an 8-bit integer to meters.
# The conversion formula is: intensity = intensity[normalized] / 0xFF * intensityScale
intensityDenormalizer @5: Float32;
# Delay in microseconds between firing
deltaTime @6: UInt16;
# Beams with a range smaller than or equal to this distance (in meters) are considered to have
# returned an invalid measurement.
invalidRangeThreshold @7: Float64;
# Beams with a range larger than or equal to this distance (in meters) are considered to not have
# hit an obstacle within the maximum possible range of the sensor.
outOfRangeThreshold @8: Float64;
}


## rigid_body_3_group

### RigidBody3GroupProto

Copy
Copied!

# A list of rigid bodies in a certain coordinate frame
struct RigidBody3GroupProto {
# List of rigid bodies
bodies @0: List(RigidBody3Proto);
# Optional names for rigid bodies
names @1: List(Text);
# Name of coordinate system reference frame
referenceFrame @2: Text;
}


### RigidBody3Proto

Copy
Copied!

# A rigid body in 3D
struct RigidBody3Proto {
# The pose of the rigid body with respect to its reference coordinate frame
refTBody @0: Pose3dProto;
# The linear velocity in the reference coordinate frame
linearVelocity @1: Vector3dProto;
# The angular velocity as a Rodrigues vector in the reference coordinate frame. This means
# the vector's direction is the axis of rotation and its length is the angular speed around
# that axis.
angularVelocity @2: Vector3dProto;
# The linear acceleration
linearAcceleration @3: Vector3dProto;
# The angular acceleration
angularAcceleration @4: Vector3dProto;
# The relative scales in x, y, z axis in the body frame; (1.0, 1.0, 1.0) represents the original
# scale.
scales @5: Vector3dProto;
}


## robot_state

### RobotStateProto

Copy
Copied!

# The robot state at that particular instance of time. The proto stores the current position and
# angle, as well as the difference between the previous values of the states (position and angle).
# Also contains the current linear and angular speed, obtained from odometry input
struct RobotStateProto {
# Current pose of robot in the world coordinate frame
worldTRobot @0: Pose2dProto;
# Current speed of body (x, y and angular)
currentSpeed @1: Vector3dProto;
# Distance travelled from last updated position
displacementSinceLastUpdate @2: Vector3dProto;
}


## segmentation_prediction

### SegmentationPredictionProto

Copy
Copied!

# Segmentation prediction proto stores the probability distribution over classes for each pixel in
# a three dimensional tensor. The probability will add up to one for each pixel.
struct SegmentationPredictionProto {
# List of all class names. The number of elements must match the third dimension of the tensor.
classNames @0: List(Text);
# Tensor with dimensions (rows * cols * classes)
prediction @1: TensorProto;
}


## state

### StateProto

Copy
Copied!

# A message used to transport states of dynamic systems. This is used closely together with the
# state gem to be found in //engine/gems/state. For example you can define the robot state using the
# compile-time types defined in that gem, and then serialize that state into a message. The
# receiver of that message can deserialize the state back into the compile-time robot state type.
struct StateProto {
# A densely packed representation of the state data as floating point numbers. The lowest rank
# indicates state element dimension, the second rank is time for a potential timeseries, and
# the third rank is for the batch index.
pack @0: TensorProto;
# The schema describing the format of the state vector
schema @1: Text;
# Alternative way to pass the data (for python)
data @2: List(Float64);
}


## superpixels

### SuperpixelLabelsProto

Copy
Copied!

# Labels for superpixels. Assigns a label to every superpixel in a related SuperpixelsProto
struct SuperpixelLabelsProto {
labels @0: List(Int32);
}


### SuperpixelsProto

Copy
Copied!

# A superpixel oversegmentation for an image. This is useful for various image segmentation methods
# or perception algorithms in general.
# # Each pixel in the image is assigned to a superpixel. For each superpixel information like the
# pixel coordinate or the color are reported. Optionally also 3D data like point and normal are
# reported.
struct SuperpixelsProto {
# 2D superpixel data based on the color image
struct Superpixel {
# The average pixel coordinate of the superpixel stored as (row, col)
pixel @0: Vector2fProto;
# The average color of the superpixel stored as RGB in the range [0,1]
color @1: Vector3fProto;
# The number of pixels in this superpixel. There might be superpixels with count = 0 which
# indicates that no pixels were assigned to them. This would also indicate that other superpixel
# and surflet data is invalid.
count @2: UInt32;
}
# Additional 3D surflet data for each superpixel defining the 3D shape of the superpixel
struct Surflet {
# 3D point coordinates of the surflet
point @0: Vector3fProto;
# 3D normal of the surflet
normal @1: Vector3fProto;
}
# An image which assignes a superpixel image to every pixel using the superpixel index. This index
# is identical to the position of the superpixel in the superpixels or surflet lists. Indices are
# stored as an unsinged 16-bit integer. Pixels which are not assigned to any superpixel are marked
# with the index 0xffff.
indices @0: ImageProto;
# 2D superpixel data for every pixel cluster
superpixels @1: List(Superpixel);
# 3D surflet data for every pixel cluster (optional)
surflets @2: List(Surflet);
}


## trajectory

### Vector3TrajectoryListProto

Copy
Copied!

# A list of trajectories with three dimensions.
# For example, one might consider the multiple trajectories for given bodies.
struct Vector3TrajectoryListProto {
trajectories @0: List(Vector3TrajectoryProto);
}


### Vector3TrajectoryProto

Copy
Copied!

# A trajectory with three dimensions.
# The trajectory is composed of a time series of positions in reference to a given frame.
# For example, one might consider the trajectory of a moving target.
# Note: The timestamps may be omitted, in which case the trajectory is equivalent to a path.
struct Vector3TrajectoryProto {
# A list of states as positions or samples.
states @0: List(Vector3dProto);
# An optional list of timestamps in seconds corresponding to the list of states or positions.
timestamps @1: List(Float64);
# The reference frame for this trajectory.
frame @2: Text;
}


## uuid

### UuidProto

Copy
Copied!

# A unique identifier following the UUID specifications
struct UuidProto {
# First 8 bytes out of 16
lower @0: UInt64;
# Second 8 bytes out of 16
upper @1: UInt64;
}


## voice_command_detection

### VoiceCommandDetectionProto

Copy
Copied!

# Message containing detected command id and list of timestamps of contributing keywords
struct VoiceCommandDetectionProto {
# Detected command id
commandId @0: UInt8;
# List of timestamps of contributing keywords to the detected command
timestamps @1: List(UInt64);
}