T1       Functions 2       Reference Model 3       I/O Data
4       Functions of AI Modules 5       I/O Data of AI Modules 6       AIW, AIMs, and JSON
7       Reference Software 8       Data Types 9     Performance Assessment

1        Functions

he Autonomous Motion Subsystem (AMS):

  1. Receives requests to reach a destination from the Human-CAV Interaction Subsystem  (HCI).
  2. Requests current Pose to Environment Sensing Subsystem (ESS).
  3. Converses with HCI (and HCI with humans) and settles on the final Route.
  4. Responds/request to specific Remote AMS requests.
  5. Requests Full Environment Descriptors subsets from select CAVs in range.
  6. Produces Full Environment Descriptors.
  7. Responds to requests for Full Environment Descriptors subsets from CAVs in range.
  8. Generates Paths (Plath Planning).
  9. Generates Trajectory (Motion Planning).
  10. Checks whether Trajectory can be implemented (Obstacle Avoidance), also considering possible Alerts.
  11. Issues Message to Motion Actuation Subsystem (MAS).
  12. Processes Message from Motion Actuation Subsystem (AMS Command Issuance).

2        Reference Model

Figure 1 gives the Autonomous Motion Subsystem Reference Model.

Figure 1 – Autonomous Motion Subsystem Reference Model

The operation of the Autonomous Motion Subsystem unfolds as follows:

  1. A human requests the Human-CAV Interaction to take them to a destination.
  2. HCI interprets request and passes the interpretation to the AMS.
  3. The AMS activates Route Planning to generate a set of Waypoints starting from the current Pose (obtained from the ESS) up to destination.
  4. The AMS
    1. Receives Basic Scene Descriptors from the ESS.
    2. Requests (subsets of) Remote CAVs’ Full Scene Descriptors and responds to similar requests from Remote CAVs.
    3. Integrates all sources of Environment Descriptors into Full Environment Descriptors
  5. The Route’s Waypoints cause the Path Planning to generate a set of Poses to reach the next Waypoint.
  6. Motion Planning generates a Trajectory to reach the next Pose in each Path.
  7. Obstacle Avoidance receives the Trajectory and checks if an Alert was received.
  8. If an Alert was received, Obstacle Avoidance detects whether the Trajectory creates a collision.
    1. If a collision is detected, Obstacle Avoidance requests a new Trajectory from Motion Planner.
    2. If no collision is detected, Obstacle Avoidance issues an AMS-MAS Message to MAS.
  9. The MAS sends an AMS-MAS Message to AMS informing about the execution of the Command.
  10. The AMS, based on the received MAS-AMS Messages, may
    1. Discontinue the execution of the earlier AMS-MAS Message.
    2. Issue a new AMS-MAS Message.
    3. Inform Obstacle Avoidance and Full Environment Description.
  11. The decision of each element of the chain may be recorded in the Decision Recording (“black box”).
  12. An authorised external system may access the AMS Recording Data.

3        I/O Data

Table 1 gives the input/output data of Autonomous Motion Subsystem.

Table 1 – I/O data of Autonomous Motion Subsystem

Input data From Comment
Basic Environment Descriptors Environment Sensing Subsystem CAV’s Environment representation from ESS.
Alert Environment Sensing Subsystem Critical information from an EST in ESS.
AMS-HCI Message Human-CAV Interaction Human commands, e.g., “take me home”.
Full Environment Descriptors Remote AMSs Other CAVs and vehicles, and roadside units.
AMS-MAS Message Motion Actuation Subsystem Message sent by the AMS to the MAS.
Ego-Remote 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”.
Full Environment Descriptors Remote AMSs To Ego CAV’s ESS and to REmote CAVs.
Ego-Remote AMS Message Remote AMSs Ego AMS to Remote AMS message.
AMS Recording Data External Device For offline analysis.

4        Functions of AI Modules

Table 2 gives the AI Modules of the Autonomous Motion Subsystem.

Table 2 – Functions of Autonomous Motion Subsystem’s AI Modules

AIM Function
Full Environment Description Creates an internal environment representation by fusing infor­mation received from ESS, Remote AMSs, and other CAV-aware entities.
Updates the CAV State.
Route Selection Planning Computes a set of possible Routes, through the road network, from the current to the target destination.
Path Selection Planning Generates a set of Paths, considering:
1. Route.
2. Full Environment Descriptors (Spatial Attitude, Road State, etc.).
4. Traffic Rules.
Motion Selection Planning Defines a Trajectory to reach a Goal using the Spatial Attitude considering:
1. CAV’s kinematic and dynamic constraints.
2. Full Environment Descriptors
3. Passengers’ comfort.
Traffic Obstacle Avoidance Checks whether Trajectory is compatible with Alert information: if it is, it passes the Trajectory to Command Issuance, if it is not, it requests a new Trajectory.
If AMS Command Issuance informs Obstacle Avoidance that AMS-MAS Message cannot be implemented, Traffic Obstacle Avoidance may issue a “discontinue previous Command” and issue a new Command.
AMS Command Issuance Instructs the MAS to execute the Trajectory considering the Environment conditions and receives MAS-AMS Messages about the execution. Based on a Message, the Road State may be communicated to Obstacle Avoidance.
AMS Decision Recording Records decisions by Route Planning, Path Planning, Motion Planning, Obstacle Avoidance, Full Environment Description, and Command Issuance.

5        I/O Data of AI Modules

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

Table 3 – Autonomous Motion Subsystem’s data

AIM Input Output
Full Environment Description – Basic Environment Descriptors
– Ego-Remote AMS Message		 – Full Environment Descriptors
Route Selection Planning – Positions
– Full Environment Descriptors
– Offline Map Object
– Selected Route		 – Route
Path Selection Planning – Route
– Full Environment Descriptors
– Offline Map Object		 – Paths
Motion Selection Planning – Paths
– Full Environment Descriptors		 – Trajectory
Traffic Obstacle Avoidance – Trajectory
– Full Environment Descriptors
– Alert
– Road State		 – Trajectory
AMS Command Issuance – Trajectory
– AMS-MAS Message		 – AMS-MAS Message
– Road State
AMS Decision Recording – Route
– Path
– Trajectory
– AMS-MAS Message
– Full Environment Descriptors		 – AMS Recording Data

6     AIW, AIMs, and JSON Metadata

AIW AIMs Name JSON
CAV-AMS Autonomous Motion Subsystem X
CAV-FEV Full Environment Description X
CAV-RSP Route Selection Planning X
CAV-PSP Path Selection Planning X
CAV-MSP Motion Selection Planning X
CAV-TOA Traffic Obstacle Avoidance X
CAV-ACI AMS-MAS Message Issuance X
CAV-ADR AMS Decision Recording X

7       Reference Software

 

8       Conformance Testing

 

9     Performance Assessment