M. Tech. Project Presentation Automatic Cruise Control System

1
M. Tech. Project Presentation
Automatic Cruise Control System

• By Rupesh Sonu Kakade
• 05323014
• Under the guidance of
• Prof. Kannan Moudgalya
• and
• Prof. Krithi Ramamritham
• Indian Institute of Technology, Bombay
• 10 July 2007

2
Overview

• Introduction
• Objectives
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

3
Introduction

• Conventional Cruise Control
• Difficulties
• 1. Useful only in sparsely populated roads
• 2. Disengagement may result in driver
• loosing control of a car.

4
Introduction

• Automatic Cruise Control (ACC) System
• Control Objectives
• 1. Follow-the-leader car
• 2. Adapt to leader velocity

5
Introduction - ACC
6
Introduction - ACC

• Safe Inter-vehicle distance Rule
• 1. Constant spacing policy Safe distance is
  independent of vehicle parameters such as maximal
  velocity, deceleration, etc.

7
Introduction - ACC

• 2. Constant time-gap policy
• Difficulties with ACC
• 1. Federal and State laws prohibits the use of
  ACC system below certain speed value.
• 2. Human driving often results in
• excessive accelerations and
• decelerations. Thus violating
• comfort specifications.

8
Introduction

• Stop-and-go scenario demands a different
  behavior from vehicles.
• Control in stop-and-go scenario
• Control Objectives
• 1. Safety Constraint Stop the vehicle before it
  reaches a critical distance, .
• 2. Comfort specification Keep the
• deceleration and jerk bounded.

9
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

10
Objectives of Project

• Design control systems for
• 1. Speed control - in conventional cruise
  control
• 2. ACC controller
• 3. Controller for stop-and-go traffic
• and
• 4. Integrate controllers on
• low-cost platform

11
Approach used

• Zones
• 1. Blue Zone Cruise control
• 2. Green Zone Automatic cruise control
• 3. Orange Zone Stop-and-go traffic control
• 4. Red Zone Safety critical zone

12
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

13
Automatic Cruise Control

• Control Objectives
• 1. Follow-the-leader car, i.e., distance error
  should be minimal. Distance error is computed
  from
• where,
• 2. Adapt to leader velocity, i.e., relative
  velocity between two vehicles should be minimal.

14
ACC

• Control Law The first time-derivative of
  distance error is computed and solved the
  following equation
• which ensures the distance error reduces to
  zero. We have

15
ACC

• The control structure is similar to PD controller
  with,
• 1. Proportional gain
• 2. Derivative gain

16
ACC Control Scheme
17
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

18
Control during stop-and-go scenario

• Control Objectives
• 1. Safety Constraint Stop the vehicle before it
  reaches a critical distance, .
• 2. Comfort specification Keep the deceleration
  and jerk values bounded for all t.
• Reference model
• Input Lead vehicle velocity and
• Output Reference distance and
• reference acceleration

19
Control during stop-and-go scenario
20
Control during stop-and-go scenario

• Reference model has twofold objectives
• 1. Reference distance computation
• 2. Reference acceleration computation
• Safety and comfort constraints

21
Control during stop-and-go scenario

• Objectives To find constraints on c and so
  that safety and comfort specifications are
  satisfied for all initial conditions and
  .
• Initial conditions are defined as
• where t 0 s, is the time when
• Orange Zone is reached.
• Solving and

22
Control during stop-and-go scenario

• where

23
Control during stop-and-go scenario

• Solving the previous expression, we have
• The maximum penetration distance is
• This gives us a lower bound on c

24
Control during stop-and-go scenario

• Next we find upper bound on c. Substitute in
  expression for reference acceleration, i.e.,
• The maximum value of reference
• breaking is computed from

25
Control during stop-and-go scenario

• Substitute in , we have

26
Control during stop-and-go scenario

• Now we consider comfort specification, i.e., jerk
  values must also be bounded. This gives us
  another upper limit on value for c.
• The maximum value of jerk is
• believed to depend on extremes of

27
Control during stop-and-go scenario

• The expression has two solutions.
• i.e., estimated lead velocity assumed
• to be zero. Therefore maximum value
• of jerk could be computed from

28
Control during stop-and-go scenario

• To proceed we assume
• i. e., negative acceleration is always
  greater than positive acceleration.
• The maximum jerk will be
• bounded as

29
Control during stop-and-go scenario

• Assuming sufficiently large for The previous
  expression
• yields another upper bound on value for c.
• C1 and c2 are associated with safety
• Whereas c3 is associated with comfort

30
Control during stop-and-go scenario

• In the Orange Zone, priority is given to safety,
  i.e.,
• Next we determine the lower bound on the value of
  .
• We use the above expression together with
• If takes the smallest value then
• c takes on the largest value.

31
Control during stop-and-go scenario
32
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

33
Results

• We implemented ACC controller on Dexter-6C. This
  platform is relatively reach in a sense that it
  has
• 1. Independent steering controller
• 2. Independent drive controller
• 3. Independent controller for white line sensing
• Our objective was to implement control system on
  a low cost platform, such as CDBOT.
• The experimental results on CDBOT are also
  presented.

34
Results

• Figure Dexter-6C, a test car

35
Results - On Dexter-6C

• Fig. Speed control loop performance Fig.
  Car-following (ACC) results

36
Results - On Dexter-6C

• Fig. Time-gap results

37
Results On CDBOT

• Inner speed control loop performance test

38
ACC Results On CDBOT
39
Results On CDBOTControl in stop-and-go scenario
40
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

41
Conclusion

• Different traffic densities is found to demand
  different behavior from vehicles.
• Controllers for longitudinal speed control of
  cars during sparsely populated road, moderate
  traffic, and stop-and-go scenarios are designed.
• Controllers were integrated on robotic platform,
  CDBOT. Also ACC controller was implemented on
  Dexter-6C.

42
Overview

• Introduction
• Objectives of project
• Automatic Cruise Control (ACC)
• Control in stop-and-go traffic
• Results
• Conclusion
• Future Improvements

43
Future Improvements

• 1. ACC controller used PD structure. Due to its
  non
• perfect tracking, jerk values are some times
  higher.
• This aspect could be improved by using advanced
• controller such as controller based on adaptive
  control
• theory.
• 2. String (or platoon) stability problem is not
  analyzed
• here.