Você está na página 1de 14

CHAPTER 3

3.0 Methodology

In this chapter the author explains how the Forward Collision Warning (FCW)
implementation on a vehicle. There are three main parts which consist of:

1. Forward Warning Collision Algorithm


2. Virtual Simulation
3. Hardware Implementation and Simulation

Figure 1 : Overall FCW Architechture

The figure above briefly shows the full architecture of FCW. First and foremost is the
algorithm needs to trigger warning and helps the driver to avoid collision from happening.
For Virtual simulation part, the author explains which software application and simulation
that has been used to implement FCW. Finally, Hardware Implementation parts explains
hardware components that has been used including sensor, controller and some
components. All parts are important for implementing FCW. Each of the parts will be
explained in this chapter.
3.1Forward Warning Collision (FCW) Algorithm

The algorithm is the main mechanism for implementing FCW. It is included with safety
threshold that has been used for the system. For the safety threshold Time to Collision
(TTC) equation has been used. The definition of TTC could be expressed as equation.

[1]

Where is the distance between the host vehicle and obstacle. In this algorithm, Time
Headway is also used for the safety time range for driving between cars. The threshold for
time headway is 5 seconds and the critical threshold time headway is 3 seconds [].

Figure 2 : Forward Warning Collision Algorithm


3.2 Virtual Simulation

In this section, the FCW simulation is denoted. Before integrating the simulation in
hardware, virtual prototype has been constructed by using MATLAB application. In
MATLAB simulation, a simple vehicle dynamic model has been used to represent the
host vehicle dynamic []. There are two situations proposed in the simulation:
1. Static Obstacle
2. Dynamic Obstacle

Thus, there are two vehicle dynamics to represent the host vehicle and frontal vehicle
which is the obstacle. The simplified vehicle dynamics modelling is as shown below:

Figure 3 : Simple Vehicle Dynamic Model


Figure 4 : Predefined Speed of Host Vehicle

In this simulation, the host vehicle speed is predefined. The host vehicle speed is
defined as shown in Figure 4. For the obstacle speed, it depends on the
aforementioned; static or dynamic condition. For the FCW algorithm, the experiment
is simulated by using MATLAB function to execute the algorithm of FCW. The
algorithm is constructed by using If Else statement in the MATLAB function. The
code of the algorithm is as shown below:

function Alert= fcn(TTC)

if TTC >5
Alert = 0;
else if TTC>3
Alert = 2;
else
Alert = 1;
end
end

end

Figure 5 : FCW algorithm code


3.4 Hardware Implementation and Simulation

3.4.1 Hardware Setup

The hardware implementation is constructed to evaluate the FCW algorithm in


real situation. To begin conducting the experiment, three main parts of the
hardware architecture are implemented:
1. Sensor
2. Control Box
3. Alert Signal

The configuration of hardware setup for this system is shown as below:

Figure 6 : Hardware setup configuration


1. Sensor

There are many type of sensor which detects obstacle. For this system, it uses
Radar as a sensor for obstacle detection. It uses Universal Medium Range Radar
(UMRR) Type 30, where it aims at medium range at wide horizontal angular
coverage.

Figure 7 : Front and Rear View of UMRR

The sensor detects obstacle and measures data that is used for developing this
FCW system. It also detects multiple obstacles simultaneously. Among the data
received from sensor are:
1. (X,Y) Position
2. Longitudinal(X) and Lateral(Y) speed.

The figure below shows the sensors field of view for FCW:

Figure 8 : Radar Field of View


2. Control Box

In this experiment, the dSPACE Micro Autobox Controller has been used as the
control box of the system. Micro Autobox Controller is a Real-Time system
controller. This control box handles the data that is collected from the sensor.
Furthermore, the control box controls the desired output signal. Figure() below
shows the Control Box Micro Autobox.

Figure 9 : Micro Autobox

2.1 Data Collection Process

To collect the data, MATLAB and ControlDesk has been used. MATLAB
software is for data manipulation while ControlDesk functions to record the data.
In MATLAB, it uses Simulink block to analyze the data that is obtained from
radar. The data that is obtained from radar can also be manipulated by using
Simulink blocks. Figure below shows the overall Simulink block diagram.
Besides that, there is another software to verify the accuracy of data collection.
The software is Driverecorder3 which is the software for the sensor. The data
collection verification of the data collected will be explained later.

From Time to Collision (TTC) equation, there are several data that need to be
collected from the sensor.

1. Distance between Host Vehicle and Obstacle


(X,Y) position
2. Relative Speed between Host Vehicle and Obstacle
I. Longitudinal (X) speed
II. Lateral (Y) speed
3. Speed of Host Vehicle The speed of vehicle is
retrieved from the CAN vehicle.
2.2 Data Validation

The data that is collected must be accurate for calculating TTC to validate
whether the data that is collected is valid by using Driverecorder3 software.
This software is the sensor software. The function of this software is to show the
obstacle that radar sensor detected. It also shows the distance and speed of
obstacle in (X,Y) coordinate form.

Figure 10 : Driverecorder 3 software user interface

All data obtain from the sensor can be validated by comparing with data
recorded from ControlDesk. The data will be shown at SensorTargetList and set
Tracking ID0. By validating the data, it ensures that the data that has been
collected are accurate. Thus this system will respond correctly when it detects
an obstacle.
2.3 Region of Investigation

From the previous explanation, sensor will detect distance in (X,Y) coordinate
form. Figure() below illustrates the co-ordinate system of the radar system. The
data are reported in Cartesian coordinates. The origin of the coordinate system
is the middle of the Host Vehicle.

Figure 11 : Cartesian coordinate system

Furthermore the radar detection angle is wide. In this experiment, the targeted
obstacle should be in the Region of Investigation (ROI). The ROI were develop
based on the typical single lane road width. By referring to the Malaysian
Federal Roads System, the standard width of road is 3.5 m. For the Maximum
distance from radar is 100 m. Thus, the ROI is formed in a rectangular shape.
In this region only the radar can detect the obstacle. In the MATLAB
simulation, the ROI has been created by using MATLAB function. The
function uses If Else statement. The code is shown as below:
Figure 12: Simulink block for Region

function Y_InRange = function X_InRange =


Yrange(u) Xrange(u)

if u>=-1.5 if u>=100
if u<=1.5 X_InRange=u;
Y_InRange=u; else
else X_InRange=0;
Y_InRange=0; end
end
else
Y_InRange=0;
end
Figure 13 : Code for Y_Range Figure 14: Code for X_range

function X = fcn(y,x)
if y==0
X=0;
else
X=x;
end

Figure 15: Code for new X Region,X


2.4 Warning State

When the obstacle is detected in the Region of Investigation (ROI), the sensor
will send data to the control box to process the threat assessment and the system
algorithm. Then, to simulate the FCW, a simplified version in Simulink block is
used. Figure() below illustrates the initial the simulation of the system. The
distance between the Host Vehicle and Front Vehicle is calculated in an initial
experiment, while the vehicle speed have no specific value and is in accordance
with the data recorded from vehicle CAN. The data that has been recorded can
be used for the calculation of Time-to-Collision equation. Thus, the TTC can be
evaluated through the system algorithm.

Figure 16 : Simulink block for Virtual Simulation

By using IF ELSE statement (Figure 5), it can be verified whether the algorithm
of the system performs as expected. The verification is very crucial because it is
the main mechanism of triggering warning state. After the verification, the
algorithm can be implemented in real-time. In the hardware implementation,
each warning state produced different types of outputs. The output of alert
signal will be described afterward.
3. Alert Signal

After implementing the algorithm in the control box, the output of the warning
state will produce alert signal. There are two types of alert signals which is LED
beacon and a sonic buzzer. Each state represents different types alert signal.

Figure 17: LED Beacon Figure 18 : 12 V Buzzer

As stated in the FCW algorithm, there are three warning states :

State Remark
1. Safety Driving No Alert
2. Warning State 1 LED Beacon On
3. Warning State 2 LED Beacon On
Buzzer On

Table 1: FCW Algorithm Output

1.4.2 Real-Time Experiment Setup

An Experiment will be done to evaluate the FCW algorithm. In this experiment,


by imitating a scenario when a car drives in a constant acceleration approaching
to a static obstacle. In this experiment, the Host Vehicle navigate with a
constant acceleration approaching to a static obstacle which is the front vehicle.
The sensor collects the data and transfer it to the control box for further
calculation. It will produce TTC value, the time taken of the host vehicle will
collide with the obstacle.

Figure 17: Scenario of the Experiment

Then, TTC Value will be formulated where a certain threshold is set. This
threshold will symbolizes the Warning States. As a result, the FCW system
should respond effectively by producing warning to the driver for vehicle
collision avoidance. The result will be explained in Chapter 4.

Você também pode gostar