1.1        Functions of Autonomous Motion Subsystem

The functions of the Autonomous Motion Subsystem (AMS) are:

  1. Receive a request to reach a destination as instructed by Human-CAV Interaction (HCI).
  2. Request current Pose to Environment Sensing Subsystem (ESS).
  3. Converse with HCI and settle on final Route.
  4. Receive Basic Environment Representation (BER) from ESS.
  5. Broadcast appropriate BER subsets to Remote AMSs.
  6. Respond to specific Remote AMS requests.
  7. Produce Full Environment Representation.
  8. Generate Paths (Plath Planner).
  9. Generate Goal and Trajectory (Motion Planner).
  10. Check whether Trajectory can be implemented (Obstacle Avoider).
  11. Issue Command to Motion Actuation Subsystem.

1.2        Reference Architecture of Autonomous Motion Subsystem

Figure 7 gives the Autonomous Motion Subsystem Reference Model.

 

This is the operation of the Reference Model:

  1. A human requests the Human-CAV Interaction to take them to a destination.
  2. HCI interprets request and passes interpretation to the AMS.
  3. The AMS activates the Route Planner to generate a set of Waypoints starting from the current Pose, obtained from the Full Environment Representation, up to the destination.
  4. The Waypoints enter the Path Planner which generates a set of Poses to reach the next Waypoint.
  5. For each Path, the Motion Planner generates a Trajectory to reach the next Pose.
  6. Obstacle Avoider receives the Trajectory and checks if an Alert was received.
  7. If an Alert was received, Obstacle Avoider checks whether the implementation of the Trajectory creates a collision.
    1. If a collision is indeed detected, Obstacle Avoider requests a new Trajectory from the Motion Planner.
    2. If no collision is detected, Obstacle Avoider issues a Command to Motion Actuation Subsystem.
  8. The Motion Actuation Subsystem sends MAS-AMS Response about the execution of the Command.
  9. The AMS, based on the MAS-AMS Responses received potentially conveying changes in the Environment, can decide to discontinue the execution of the earlier Command and issue another AMS-MAS Message instead.
  10. The decision of each element of the said chain may be recorded in the Decision Recorder (“black box”).

 

Figure 7 – Autonomous Motion Subsystem Reference Model

1.3        I/O Data of Autonomous Motion Subsystem

Table 12 gives the input/output data of Autonomous Motion Subsystem.

 

Table 12 – I/O data of Autonomous Motion Subsystem

Input data From Comment
Basic Environment Representation Environment Sensing Subsystem CAV’s Environment representation.
Alert Environment Sensing Subsystem Critical information from an EST in ESS.
HCI-AMS Message Human-CAV Interaction Human commands, e.g., “take me home”
Environment Representation Remote AMSs Other CAVs and vehicles, and roadside units.
MAS-AMS Message Motion Actuation Subsystem Message sent by the AMS to the MAS.
Remote-Ego AMS Message Remote AMS Remote AMS to Ego AMS message
Output data To Comment
AMS-HCI Message Human-CAV Interaction AMS’s message to HCI-AMS.
AMS-MAS Message Motion Actuation Subsystem Message to MAS, e.g., “in 5s assume a given Spatial Attitude”.
Environment Representation Remote AMSs For information to other CAVs
Ego-Remote AMS Message Remote AMS Ego AMS to Remote AMS message

1.4        Functions of Autonomous Motion Subsystem’s AI Modules

Table 13 gives the AI Modules of the Autonomous Motion Subsystem.

 

Table 13 – Functions of Autonomous Motion Subsystem’s AI Modules

AIM Function
Environment Representation Fusion Creates an internal representation of the Environment by fusing infor­mation from itself, Remote AMSs, and other CAV-aware entities.
Route Selection Planner
Path Selection Planner Generates a set of Paths, considering:

1.      Route.

2.      Spatial Attitude.

3.      Full Environment Representation.

4.      Traffic Rules.
Motion Selection Planner Defines a Goal and a Trajectory to reach the Goal using the Spatial Attitude

1.      Satisfying the CAV’s kinematic and dynamic constraints

2.      Considering passengers’ comfort.
Traffic Obstacle Avoider Checks whether the Trajectory is compatible with any Alert information. If it is, it passes the Trajectory to the Command Issuer. If it is not, it requests a new Trajectory. If Command Issuer informs Obstacle Avoider that AMS-MAS Message cannot be implemented, Obstacle Avoider may issue a “discontinue previous Command” and forward to the appropriate upstream AIM, possibly up to the Route Planner. This may decide to communicate the Road State to the Human-CAV Interaction Subsystem.
MAS Command Issuer Instructs the MAS to execute the Trajectory considering the Environment conditions and receives MAS-AMS Responses about the execution. Based on the Response the Road State may be communicated to the Obstacle Avoider.
AMS Decision Recorder Records decisions by Route Planner, Path Planner, Motion Planner, Obstacle Avoider, Command Issuer.

1.5        I/O Data of Autonomous Motion Subsystem’s AI Modules

Table 14 gives, for each AIM (1st column), the input data (2nd column) and the output data (3rd column) of Autonomous Motion Subsystem.

 

Table 14 – Autonomous Motion Subsystem’s data

CAV/AIM Input Output
Environment Representation Fusion Basic Environment Representations

from ESS or Remote AMSs

Other data from Remote AMSs

 Full Environment Representation
Route Selection Planner Poses

Full Environment Representation

Offline Map Data

Road State

Selected Route

 Route
Path Selection Planner Route

Full Environment Representation

Offline maps

Road State

 Set of Paths
Motion Selection Planner Set of Paths

Full Environment Representation

Road State

 Trajectory
Traffic Obstacle Avoider Trajectory

Full Environment Representation

Alert

Road State

 Trajectory
MAS Command Issuer Trajectory

MAS-AMS Message

 AMS-MAS Message
AMS Decision Recorder Route

Path

Trajectory

AMS-MAS Message

MAS-AMS Message

 Recorded Data