This Node is responsible for predicting environment objects' motion and sharing those predictions with the rest of CARMA Platform. It serves three main functions:

Convert incoming BSM, PSM, and mobility path data into ExternalObjects
Predict incoming objects' motion using either the constant velocity (CV) or constant turn-rate and velocity (CTRV) model
Queue and synchronize incoming object information so that it gets published all at once at a specific frequency

The following table lists the motion model used for each ExternalObject semantic category

Object type	Motion model
`UNKNOWN`	CTRV
`MOTORCYCLE`	CTRV
`SMALL_VEHICLE`	CTRV
`LARGE_VEHICLE`	CTRV
`PEDESTRIAN`	CV
default	CV

‍[!IMPORTANT]\ This Node requires a valid georeference to do the motion predictions. Until it receives one on the georeference topic, it will not publish any.

‍[!IMPORTANT]\ This Node will only publish motion predictions when it receives external object information on the external_objects topic. Messages on the incoming_* topics will be queued until then.

Subscriptions

Topic	Message Type	Description
`external_objects`	`carma_perception_msgs::msg::ExternalObjectList`	Locally-detected objects
`incoming_mobility_path`	`carma_v2x_msgs::msg::MobilityPath`	Self-reported planned motion paths from nearby vehicles
`incoming_bsm`	`carma_v2x_msgs::msg::BSM`	Incoming basic safety messages
`incoming_psm`	`carma_v2x_msgs::msg::PSM`	Incoming personal safety messages
`georeference`	`std_msgs::msg::String`	Georeference point for projecting WGS-84 coordinates to a plane. Assumed and required to be a valid PROJ string.

Publishers

Topic	Message Type	Frequency	Description
`external_object_predictions`	`carma_perception_msgs::msg::ExternalObjectList`	Subscription-driven	This Node will publish predictions only when it receives a message from the `external_objects` topic. Incoming mobility paths, BSMs, and PSMs will be queued until then.

Parameters

Topic	Data Type	Default Value	Required	Read Only	Description
`prediction_time_step`	`double`	`0.1`	Yes	No	Motion prediction time step (in seconds)
`mobility_path_time_step`	`double`	`0.1`	Yes	No	Time between received mobility path predicted states (in seconds)
`prediction_period`	`double`	`2.0`	Yes	No	Prediction horizon (in seconds)
`cv_x_accel_noise`	`double`	`9.0`	Yes	No	CV Model X-Axis Acceleration Noise
`cv_y_accel_noise`	`double`	`9.0`	Yes	No	CV Model Y-Axis Acceleration Noise
`prediction_process_noise_max`	`double`	`1000.0`	Yes	No	Maximum expected process noise; used for mapping noise to confidence in [0,1] range
`prediction_confidence_drop_rate`	`double`	`0.95`	Yes	No	Percentage of initial confidence to propagate to next time step
`enable_bsm_processing`	`bool`	`false`	Yes	No	If `true` then BSM messages will be converted to `ExternalObjects`. If other object sources are enabled, they will be synchronized but no fusion will occur (objects may be duplicated)
`enable_psm_processing`	`bool`	`false`	Yes	No	If `true` then PSM messages will be converted to `ExternalObjects`. If other object sources are enabled, they will be synchronized but no fusion will occur (objects may be duplicated)
`enable_mobility_path_processing`	`bool`	`true`	Yes	No	If `true` then MobilityPath messages will be converted to `ExternalObjects`. If other object sources are enabled, they will be synchronized but no fusion will occur (objects may be duplicated)
`enable_sensor_processing`	`bool`	`false`	Yes	No	If `true` then `ExternalObjects` generated from sensor data will be processed. If other object sources are enabled, they will be synchronized but no fusion will occur (objects may be duplicated)

Services

This Node does not provide services.

Actions

This Node does not provide actions.