motion_computation_worker.cpp

Go to the documentation of this file.

// Copyright 2019-2023 Leidos
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
 
#include <wgs84_utils/proj_tools.h>
 
#include <memory>
#include <string>
 
#include <motion_computation/message_conversions.hpp>
#include <motion_computation/motion_computation_worker.hpp>
 
namespace motion_computation
{
MotionComputationWorker::MotionComputationWorker(
  const PublishObjectCallback & obj_pub,
  rclcpp::node_interfaces::NodeLoggingInterface::SharedPtr logger,
  rclcpp::node_interfaces::NodeClockInterface::SharedPtr node_clock)
: obj_pub_(obj_pub), logger_(logger), node_clock_(node_clock)
{
}
 
void MotionComputationWorker::predictionLogic(
  carma_perception_msgs::msg::ExternalObjectList::UniquePtr obj_list)
{
  carma_perception_msgs::msg::ExternalObjectList sensor_list;
  sensor_list.header = obj_list->header;
 
  if (!obj_list || obj_list->objects.empty()) {
    return;
  }
 
  for (auto obj : obj_list->objects) {
    // Header contains the frame rest of the fields will use
    // obj.header = obj_list.objects[i].header;
 
    // Object id. Matching ids on a topic should refer to the same object within
    // some time period, expanded
    // obj.id = obj_list.objects[i].id;
 
    // Update the object type and generate predictions using CV or CTRV vehicle models.
    // If the object is a bicycle or motor vehicle use CTRV otherwise use CV.
 
    bool use_ctrv_model;
 
    if (obj.object_type == obj.UNKNOWN) {
      use_ctrv_model = enable_ctrv_for_unknown_obj_;
    } else if (obj.object_type == obj.MOTORCYCLE) {
      use_ctrv_model = enable_ctrv_for_motorcycle_obj_;
    } else if (obj.object_type == obj.SMALL_VEHICLE) {
      use_ctrv_model = enable_ctrv_for_small_vehicle_obj_;
    } else if (obj.object_type == obj.LARGE_VEHICLE) {
      use_ctrv_model = enable_ctrv_for_large_vehicle_obj_;
    } else if (obj.object_type == obj.PEDESTRIAN) {
      use_ctrv_model = enable_ctrv_for_pedestrian_obj_;
    } else {
      obj.object_type = obj.UNKNOWN;
      use_ctrv_model = enable_ctrv_for_unknown_obj_;
    }  // end if-else
 
    if (use_ctrv_model == true) {
      obj.predictions = motion_predict::ctrv::predictPeriod(
        obj, prediction_time_step_, prediction_period_, prediction_process_noise_max_,
        prediction_confidence_drop_rate_);
    } else {
      obj.predictions = motion_predict::cv::predictPeriod(
        obj, prediction_time_step_, prediction_period_, cv_x_accel_noise_, cv_y_accel_noise_,
        prediction_process_noise_max_, prediction_confidence_drop_rate_);
    }
 
    sensor_list.objects.emplace_back(obj);
  }  // end for-loop
 
  // Synchronize all data to the current sensor data timestamp
  carma_perception_msgs::msg::ExternalObjectList synchronization_base_objects;
  synchronization_base_objects.header =
    sensor_list.header;  // Use the current sensing stamp as the sync point even if sensor data is
                         // not used
 
  if (enable_sensor_processing_) {
    // If using sensor data add it to the base synchronization list since it
    // already is at the desired time
 
    synchronization_base_objects.objects = sensor_list.objects;
 
  } else if (enable_bsm_processing_ || enable_psm_processing_ || enable_mobility_path_processing_) {
    // Since we use the new sensor data as the sync point we will still be
    // synchronizing but to an empty list
    synchronization_base_objects.objects.clear();
 
  } else {
    RCLCPP_WARN_STREAM(
      logger_->get_logger(),
      "Not configured to publish any data publishing empty "
      "object list. Operating like this is NOT "
      "advised.");
 
    obj_pub_(synchronization_base_objects);
    bsm_list_.objects.clear();
    bsm_obj_id_map_.clear();
    psm_list_.objects.clear();
    psm_obj_id_map_.clear();
    mobility_path_list_.objects.clear();
    mobility_path_obj_id_map_.clear();
 
    return;
  }
 
  // Start synchronizing all the enabled data streams
  if (enable_bsm_processing_) {
    synchronization_base_objects = synchronizeAndAppend(synchronization_base_objects, bsm_list_);
  }
 
  if (enable_psm_processing_) {
    synchronization_base_objects = synchronizeAndAppend(synchronization_base_objects, psm_list_);
  }
 
  if (enable_mobility_path_processing_) {
    synchronization_base_objects =
      synchronizeAndAppend(synchronization_base_objects, mobility_path_list_);
  }
 
  obj_pub_(synchronization_base_objects);
  // Clear msg queue since it is published
  mobility_path_list_.objects.clear();
  mobility_path_obj_id_map_.clear();
  bsm_list_.objects.clear();
  bsm_obj_id_map_.clear();
  psm_list_.objects.clear();
  psm_obj_id_map_.clear();
}
 
void MotionComputationWorker::georeferenceCallback(const std_msgs::msg::String::UniquePtr msg)
{
  // Build projector from proj string
  if (georeference_ != msg->data)
  {
    georeference_ = msg->data;
    map_projector_ = std::make_shared<lanelet::projection::LocalFrameProjector>(msg->data.c_str());
 
    std::string axis =
      wgs84_utils::proj_tools::getAxisFromProjString(msg->data);  // Extract axis for orientation
 
    RCLCPP_INFO_STREAM(logger_->get_logger(), "Extracted Axis: " << axis);
 
    ned_in_map_rotation_ =
      wgs84_utils::proj_tools::getRotationOfNEDFromProjAxis(axis);  // Extract map rotation
 
    RCLCPP_DEBUG_STREAM(
      logger_->get_logger(), "Extracted NED in Map Rotation (x,y,z,w) : ( "
                              << ned_in_map_rotation_.getX() << ", " << ned_in_map_rotation_.getY()
                              << ", " << ned_in_map_rotation_.getZ() << ", "
                              << ned_in_map_rotation_.getW());
  }
}
 
void MotionComputationWorker::setPredictionTimeStep(double time_step)
{
  prediction_time_step_ = time_step;
}
 
void MotionComputationWorker::setPredictionPeriod(double period) { prediction_period_ = period; }
 
void MotionComputationWorker::setXAccelerationNoise(double noise) { cv_x_accel_noise_ = noise; }
 
void MotionComputationWorker::setYAccelerationNoise(double noise) { cv_y_accel_noise_ = noise; }
 
void MotionComputationWorker::setProcessNoiseMax(double noise_max)
{
  prediction_process_noise_max_ = noise_max;
}
 
void MotionComputationWorker::setConfidenceDropRate(double drop_rate)
{
  prediction_confidence_drop_rate_ = drop_rate;
}
 
void MotionComputationWorker::setDetectionInputFlags(
  bool enable_sensor_processing, bool enable_bsm_processing, bool enable_psm_processing,
  bool enable_mobility_path_processing)
{
  enable_sensor_processing_ = enable_sensor_processing;
  enable_bsm_processing_ = enable_bsm_processing;
  enable_psm_processing_ = enable_psm_processing;
  enable_mobility_path_processing_ = enable_mobility_path_processing;
}
 
void MotionComputationWorker::setDetectionMotionModelFlags(
  bool enable_ctrv_for_unknown_obj, bool enable_ctrv_for_motorcycle_obj,
  bool enable_ctrv_for_small_vehicle_obj, bool enable_ctrv_for_large_vehicle_obj,
  bool enable_ctrv_for_pedestrian_obj)
{
  enable_ctrv_for_unknown_obj_ = enable_ctrv_for_unknown_obj;
  enable_ctrv_for_motorcycle_obj_ = enable_ctrv_for_motorcycle_obj;
  enable_ctrv_for_small_vehicle_obj_ = enable_ctrv_for_small_vehicle_obj;
  enable_ctrv_for_large_vehicle_obj_ = enable_ctrv_for_large_vehicle_obj;
  enable_ctrv_for_pedestrian_obj_ = enable_ctrv_for_pedestrian_obj;
}
 
void MotionComputationWorker::mobilityPathCallback(
  const carma_v2x_msgs::msg::MobilityPath::UniquePtr msg)
{
  if (!map_projector_) {
    RCLCPP_ERROR(
      logger_->get_logger(), "Map projection not available yet so ignoring MobilityPath messages");
    return;
  }
 
  if (!enable_mobility_path_processing_) {
    RCLCPP_WARN(
      logger_->get_logger(),
      "enable_mobility_path_processing is false so ignoring MobilityPath messages");
    return;
  }
 
  carma_perception_msgs::msg::ExternalObject obj_msg;
  conversion::convert(*msg, obj_msg, *map_projector_);
 
  // Check if this mobility path is from an object already being queded.
  // If so then update the existing object, if not add it to the queue
  if (mobility_path_obj_id_map_.find(obj_msg.id) != mobility_path_obj_id_map_.end()) {
    mobility_path_list_.objects[mobility_path_obj_id_map_[obj_msg.id]] = obj_msg;
 
  } else {
    // Add the new object to the queue and save its index
    mobility_path_obj_id_map_[obj_msg.id] = mobility_path_list_.objects.size();
    mobility_path_list_.objects.push_back(obj_msg);
  }
}
 
void MotionComputationWorker::psmCallback(const carma_v2x_msgs::msg::PSM::UniquePtr msg)
{
  if (!map_projector_) {
    RCLCPP_DEBUG_STREAM(
      logger_->get_logger(), "Map projection not available yet so ignoring PSM messages");
    return;
  }
 
  if (!enable_psm_processing_) {
    RCLCPP_DEBUG_STREAM(
      logger_->get_logger(), "enable_psm_processing is false so ignoring PSM messages");
    return;
  }
 
  carma_perception_msgs::msg::ExternalObject obj_msg;
  conversion::convert(
    *msg, obj_msg, map_frame_id_, prediction_period_, prediction_time_step_, *map_projector_,
    ned_in_map_rotation_, node_clock_);
 
  // Check if this psm is from an object already being queded.
  // If so then update the existing object, if not add it to the queue
  if (psm_obj_id_map_.find(obj_msg.id) != psm_obj_id_map_.end()) {
    psm_list_.objects[psm_obj_id_map_[obj_msg.id]] = obj_msg;
 
  } else {
    // Add the new object to the queue and save its index
    psm_obj_id_map_[obj_msg.id] = psm_list_.objects.size();
    psm_list_.objects.push_back(obj_msg);
  }
}
 
void MotionComputationWorker::bsmCallback(const carma_v2x_msgs::msg::BSM::UniquePtr msg)
{
  if (!map_projector_) {
    RCLCPP_DEBUG_STREAM(
      logger_->get_logger(), "Map projection not available yet so ignoring PSM messages");
    return;
  }
 
  if (!enable_bsm_processing_) {
    RCLCPP_DEBUG_STREAM(
      logger_->get_logger(), "enable_bsm_processing is false so ignoring BSM messages");
    return;
  }
 
  carma_perception_msgs::msg::ExternalObject obj_msg;
  conversion::convert(
    *msg, obj_msg, map_frame_id_, prediction_period_, prediction_time_step_, *map_projector_,
    ned_in_map_rotation_);
 
  // Check if this bsm is from an object already being queded.
  // If so then update the existing object, if not add it to the queue
  if (bsm_obj_id_map_.find(obj_msg.id) != bsm_obj_id_map_.end()) {
    bsm_list_.objects[bsm_obj_id_map_[obj_msg.id]] = obj_msg;
 
  } else {
    // Add the new object to the queue and save its index
    bsm_obj_id_map_[obj_msg.id] = bsm_list_.objects.size();
    bsm_list_.objects.push_back(obj_msg);
  }
}
 
carma_perception_msgs::msg::ExternalObjectList MotionComputationWorker::synchronizeAndAppend(
  const carma_perception_msgs::msg::ExternalObjectList & base_objects,
  carma_perception_msgs::msg::ExternalObjectList new_objects) const
{
  carma_perception_msgs::msg::ExternalObjectList output_list;
  output_list.header = base_objects.header;
  output_list.objects.reserve(base_objects.objects.size() + new_objects.objects.size());
 
  // Compare time_stamps of first elements of each list as they are guaranteed to be the
  // earliest of the respective lists
 
  for (auto & obj : new_objects.objects) {
    // interpolate and match timesteps
    obj = matchAndInterpolateTimeStamp(obj, rclcpp::Time(base_objects.header.stamp));
  }
 
  output_list.objects.insert(
    output_list.objects.begin(), base_objects.objects.begin(), base_objects.objects.end());
  output_list.objects.insert(
    output_list.objects.end(), new_objects.objects.begin(), new_objects.objects.end());
  return output_list;
}
 
carma_perception_msgs::msg::ExternalObject MotionComputationWorker::matchAndInterpolateTimeStamp(
  carma_perception_msgs::msg::ExternalObject path, const rclcpp::Time & time_to_match) const
{
  carma_perception_msgs::msg::ExternalObject output = path;
  // empty predictions
  output.predictions = {};
 
  // add the first point to start of the predictions to easily loop over
  carma_perception_msgs::msg::PredictedState prev_state;
  prev_state.header.stamp = output.header.stamp;
  prev_state.predicted_position.orientation = output.pose.pose.orientation;
  prev_state.predicted_velocity = output.velocity.twist;
  prev_state.predicted_position.position.x = output.pose.pose.position.x;
  prev_state.predicted_position.position.y = output.pose.pose.position.y;
  prev_state.predicted_position.position.z = output.pose.pose.position.z;
  path.predictions.insert(path.predictions.begin(), prev_state);
 
  rclcpp::Time curr_time_to_match = time_to_match;
 
  // because of this logic, we would not encounter mobility path
  // that starts later than the time we are trying to match (which is starting
  // time of sensed objects)
  bool is_first_point = true;
  carma_perception_msgs::msg::PredictedState new_state;
  for (auto const & curr_state : path.predictions) {
    if (curr_time_to_match > curr_state.header.stamp) {
      prev_state = curr_state;
      continue;
    }
 
    if (is_first_point)  // we store in the body if it is the first point, not predictions
    {
      // reaching here means curr_state starts later than the time we are trying to match
      // copy old unchanged parts
      new_state.header.stamp = curr_time_to_match;
      new_state.predicted_velocity = prev_state.predicted_velocity;
      new_state.predicted_position.orientation = prev_state.predicted_position.orientation;
 
      // interpolate position
      rclcpp::Duration delta_t = rclcpp::Time(curr_state.header.stamp) - curr_time_to_match;
      rclcpp::Duration pred_delta_t =
        rclcpp::Time(curr_state.header.stamp) - rclcpp::Time(prev_state.header.stamp);
      double ratio;
      if (pred_delta_t.seconds() < 0.00000001) {  // Divide by zero check
        // This can only happen if effectively all 3 points are on top of each other
        // which is extremely unlikely
        ratio = 0.0;
      } else {
        ratio = delta_t.seconds() / pred_delta_t.seconds();
      }
 
      double delta_x =
        curr_state.predicted_position.position.x - prev_state.predicted_position.position.x;
      double delta_y =
        curr_state.predicted_position.position.y - prev_state.predicted_position.position.y;
      double delta_z =
        curr_state.predicted_position.position.z - prev_state.predicted_position.position.z;
      // we are "stepping back in time" to match the position
      new_state.predicted_position.position.x =
        curr_state.predicted_position.position.x - delta_x * ratio;
      new_state.predicted_position.position.y =
        curr_state.predicted_position.position.y - delta_y * ratio;
      new_state.predicted_position.position.z =
        curr_state.predicted_position.position.z - delta_z * ratio;
 
      output.header.stamp = curr_time_to_match;
      output.pose.pose.orientation = new_state.predicted_position.orientation;
      output.velocity.twist = new_state.predicted_velocity;
      output.pose.pose.position.x = new_state.predicted_position.position.x;
      output.pose.pose.position.y = new_state.predicted_position.position.y;
      output.pose.pose.position.z = new_state.predicted_position.position.z;
      is_first_point = false;
 
    } else {
      output.predictions.push_back(curr_state);
    }
  }
 
  return output;
}
 
}  // namespace motion_computation