Optimized Vehicle Control Communication Interaction During Platoon Maneuvers in an AHS Environment N

1
Optimized Vehicle Control /Communication
Interaction During Platoon Maneuversin an AHS
EnvironmentNovember 5th, 1998

• Prof. Karl Hedrick
• Yu-Han Chen and Sonia Mahal
• University of California, Berkeley

2
Presentation Outline

• Research Goals
• Background and Motivation
• Work Completed to Date
• Work in Progress

3
Research Goals

• Design of the longitudinal coordination protocols
• Optimize the control algorithms to work
  effectively with a communication system
• Define the communication requirements of both the
  protocol and control algorithms

4
Background

• AHS Environment consists of
• platoons
• free agents (platoon of size 1)
• Platoon is a group of tightly spaced vehicles
  traveling at relative high speeds
• Platoon Maneuvers
• join
• split
• lane change

5
Control/Communication

• Control Standpoint
• control laws have been developed and implemented
• these control laws were successfully executed at
  the AHS Demo 97
• Communication Standpoint
• messages required for inter-platoon controls have
  been successfully delivered

6
Control/Communication (cont.)

• Control
• control laws have not been tested
• for their robustness with regards to packet
  losses
• for their performance during switching
• Communication
• platoon maneuvers have not be initiated through
  communications
• specifications for quality of service required to
  initiate and execute a platoon maneuver have not
  been determined

7
Background Motivation

• Various communication architectures have been
  used to transmit data between vehicles in a
  platoon to maintain stable vehicle spacing
• Knowledge from these existing architectures can
  be used to derive achievable specifications for
  platoon maneuvers

8
Definition of AHS Communication Messages

• Control Messages contain information required to
  maintain stable inter-vehicle spacing
• preceding vehicles velocity and acceleration
• lead cars velocity and acceleration
• preceding vehicles magnet count
• Command Messages contain information regarding a
  particular maneuver or an emergency
• request to initiate a platoon split
• handshaking messages
• abort maneuver

9
Background Motivation

• Control messages are time critical
• Specifications for time to handle control
  messages has been crucial
• Need for command messages has been acknowledged
  but not examined in detail
• Command messages are not time critical
• Reasonable bounds on latency must be specified
• One of the goals of this project is to explore
  what would be a reasonable time to complete
  platoon maneuvers

10
Steps to initiate a platoon split

• Vehicle X sends a request to split
• Platoon Leader receives request
• Platoon sends out a prepare to split message
• the request is granted only if no other maneuvers
  are in progress
• Each follower receivers this message and responds
  with an acknowledgement
• Platoon Leader issues a message to execute the
  split maneuver once it receives all
  acknowledgements

11
Existing Communication Systems

• Infrared - line of sight
• WaveLAN - token bus, pt-pt, broadcast

L
2
3
N
12
Existing Communication Systems (cont.)

• Utilicom/Hughes - pure TDMA, 3 channels
• control messages are sent every 20 ms
• command messages are sent every 60 ms

13
Existing Communication Systems (cont.)

• MPI
• LAN WAN
• front to back pt-pt,
  broadcast, CSMA
• control messages command
  messages

L
L
2
2
FAP
3
3
N
N
14
Time comparison between existing communication
systems to initiate a platoon split maneuver

• Infrared 3x(n-1) 8.5ms
• WaveLAN 3t 60ms
• Utilicom/Hughes 2s 6t 120ms
• MPI 2d t d(n-1) d(n-1) t 2dn
  154ms
• t time to send control messages
• s time to send command messages
• d time to send message to roadside radio
• n size of platoon
• x time to send data between vehicles

15
Work Done to Date

• Hardware limitations to do platoon maneuvers from
  existing communication systems have been explored
• Protocols to coordinate the join and split
  platoon maneuver has been developed

16
Work Done to Date

• Finite State Machines for longitudinal platoon
  maneuvers have been designed and verified
• logically tested using Cospan with no packet
  losses
• fault tree analysis schemes have been used to
  show all possible outcomes if a packet loss
  occurred
• FSM have been coded in C for implementation
• Success and Abort branches have been tested
  statically using the test vehicles
• Control laws have been simulated using CombSim

17
Work In Progress

• SmartAHS Simulations using the SHIFT Programming
  Language.
• code platoon maneuvers and simulate their
  robustness with regard to communication packet
  losses
• Dynamic Testing of Platoon Maneuvers
• four car test using WaveLAN radios
• Analysis of Control Laws
• robustness towards packet losses
• performance during switching

18
Work In Progress

• To estimate how many maneuvers might occur on an
  actual highway
• the amount of data being transmitted
• predict packet losses that may occur
• type of time delays that might be encountered to
  execute a maneuver
• check if these time delays are reasonable
• overall the logistics of executing these
  maneuvers