DSP IMPLEMENTATION OF A-CONTROL ALGORITHM FOR A FORWARDER CRANE

(1)

DSP IMPLEMENTATION OF A-CONTROL

ALGORITHM FOR A FORWARDER CRANE

Naga Praveen Parchuru

Jagadeesh Thati

This thesis is presented as part of Degree of

Master of Science in Electrical Engineering

Blekinge Institute of Technology

December 2009

Blekinge Institute of Technology School of Engineering Department of Electrical Engineering

(2)

Symbols

𝜽

𝟏

The swivel arm angle

𝜽

_𝟐

The lift arm angle

𝜽

_𝟑

The elbow arm angle

𝒅

𝟒

Length of prismatic joint

𝜽

_𝟏

,

𝜽

_𝟐

Angular velocities of both the links

𝒅

_𝟒

Joint velocity of the extension link

J(q) Jacobian function

v Velocity function

𝑶

_𝟒

The crane tip position

r, z Cartesian coordinates in r and z coordinates

𝒅

_𝟒𝒅

The desired prismatic joint velocity

𝜽

_𝟐𝒎

, 𝜽

_𝟑𝒎

, 𝒅

_𝟒𝒎

Maximum joint velocities of the joints

𝒌 Gain parameter which is adjustable

𝒅

_𝟒𝒄

The Center position of the prismatic joint

𝒅

_𝟒𝒆

The End position of the prismatic joint

T The Transformation matrix

(3)

Foreword/Acknowledgement

First, we thank our respected Supervisors, Mr. JONAS LINDHOLM, Mr. ANDERS

HULTGREN and Mr. BJORN NILSSON who guided us in right way to accomplish this

project in time. Whole hearted thanks to Dasa Control Systems AB, Vaxjo University,

Blekinge Institute of Technology Sweden as well as Rottne Industry AB, for providing us

such a good platform for carrying out the course and project in a satisfactory manner.

Finally, our gratitude to our all friends and teachers, those who gave help and assistance

needed. God bless all of them and we express our deep gratitude.

Naga Praveen Parchuru

Jagadeesh Thati

(4)

1. INTRODUCTION

1.1 Forwarder Introduction:

In the design of automated heavy vehicles, especially in the areas of control systems and robotics, the

engineers and researchers meet exiting challenges. The field of automation has gained its importance making operation of heavy vehicles much easier e.g., in the design of forwarders. Forwarders are used for handling timber in forest. Still there is much to do in order to improve the forwarder operators working condition.

The main task for a forwarder operator is to control the crane collecting logs from the ground and loading the forwarder wagon, or unloading the forwarder and putting the logs in piles. While controlling the crane the operator has to handle several joy stick functions. The main objective for this work is to improve the working condition for the forwarder operator, by means of partly automate the crane. This thesis project is about implementation of an algorithm for automation of one crane link i.e., the prismatic link.

1.2 Thesis Project:

This thesis mainly concentrates on implementation of the algorithm of the automatic control of the prismatic link of a crane forwarder. The implementation is made in a Texas DSP as a part of the control system for the forwarder. The algorithm is called “Software 1A”. It should be implemented using C/C++ programming language. The algorithm is developed in a research project* within forwarder crane control at Linneus University and Blekinge Institute of Technology. The algorithm is tested by use of a dSpace system controlling a laboratory crane. The code is available as Matlab and Simulink code.This code needs to be made compatible to DSP code that is needed to be implemented in the Code Composer studio (CCS). With the help of dasa5 system equipment which is provided by Dasa Company, we build the user interface for the desired algorithm developed in Code Composer Studio (CCS) for the operation of the crane links. The developed algorithm should first be tested in the Dasa lab and then it should be tested in the crane laboratory at the Linneus University in Vaxjo.

1.3 Current Ongoing Projects on Cranes:

Various current projects are mainly concentrated on improved control algorithms for the cranes, which make

use of automation as the means of driving force. This helps to reduce the time consumption and speeds up the operation and makes itself operator friendly. Research projects within control of forwarder cranes are taking place at several places, e.g. Umea University[1] and Swedish Royal Technical University[2].

(8)

1.4 Problem Definition:

The thesis task is to implement the algorithm for the automatic extension link into a proper Dasa computer

system in order to control a forwarder crane. The implemented code should be tested at the laboratory crane at Linneus University, Vaxjo.

1.5 Goals:

The main objective of this thesis is to develop Software 1A in C/C++ programming language, which is obtained from the Matlab code. This developed Software 1A should be able to be interfaced with the Code Composer Studio of the Texas Instruments along with the dasa5 system which able to run the algorithm.

The goals of this thesis can be divided into mainly three parts. Firstly, we should develop the lab testing environments at dasa and verify the results using the Visual studio C++.Secondly, we need to verify the developed algorithm in the crane lab in the practical environment. Lastly, we need to design the parameters which could suite the real crane environment with a suitable forwarder. The forwarder will then be used for the tests of the implemented algorithm.

1.6 Tools Required:

Implementation of our project urges to work on specific software which is best suited for the target system for its efficient functioning. We used the C-code in Visual Studio C++ software programming language that could be made compatible for the Code Composer Studio (CCS) of Texas Instruments to execute the result.

Digital Signal Processing applications can also be performed with the use of software that is provided by Texas Instruments. User friendly environment is provided by the Code Composer Studio where in which the C programming is used. The DSP processors of family TMS320C6712 is of in specific. The dasa5 software which is the software provided by dasa company on its own, is also made used in implementing the crane tip control. The graphical user interface for the crane control is implemented in the Dasa5 software typically. The GUI environment with the proper assignment of parameters those which can be also configured using the dasa software. The parameters of choice are used in the designing of the code in the Code Composer Studio (CCS) and are interfaced with the dasa5 software for the functioning of the crane from the operators’ point of view. The parameters can be such as buttons, voltmeters, ammeters, power meters, and indicators etc which are available from the dasa developer software. These can be interfaced with the Code Composer Studio end with developing the code for automatic crane tip control.

(9)

1.7 Contents of Thesis:

In the following chapters of the thesis we have discussed various topics and finally reached the goals. In chapter 2, the main concentration is about the background of software that was used, in order to implement the

algorithm. The kinematics of the crane is also discussed in this chapter which may include the working of the crane, forwarder kinematics and about the present forwarders that were in use in forestry purposes. Thus this chapter briefs the core idea to proceed further in accomplishing the goals.

In chapter 3, is mainly concerned on the prismatic link algorithm development. This chapter also includes the parameters that were chosen to be configurable.

In chapter 4 the explanation of RK-62 crane parameters is carried out. Specifically the various parameters of the crane links are discussed.

In chapter 5 a more detail discussion is given about dasa equipment and all their functionality.

In chapter 6 the clear idea about the crane lab environment that could give us the clear picture of the physical set up of a real crane which was provided by the Rottne Industry AB at the Vaxjo University. Detail scenario of the physical connections of the real crane is studied. Replacing the old dSpace system with the new dasa equipment is also studied.

In chapter 7 the Agility MCS tool is discussed in detail. Thus the knowledge about the efficiency, the method of conversion and the utilities is discussed in detail.

In chapter 8 the conclusion of the project of the automatic extension link of the crane is given ,chapter 9 we list out the references and in chapter 10 discuss the graphical representations UML representations of the software implemented in this thesis.

(10)

(11)

2. BACKGROUND

2.1 Code Composer Studio (CCS):

Code Composer Studio (CCS), provides the necessary software support tools. It gives an integrated development environment, bringing together the C compiler, assembler, linker, and debugger. It has graphical capabilities and support real-time debugging capabilities. Thus the Code Composer Studio is an easy to use software tool to build and debug programs. The code composer studio includes the tools for code generation which is its explicit characteristics. Thus we develop the C code in this environment and are interfaced to other software simultaneously.

2.2 d5 Developer Software:

The d5 developer is a combination of various developing tools for developing, configuring, testing, and troubleshooting purposes. It comprises of five sub functions software, which are d5DCL, d5 Designer, d5Access, d5 Analyzer and d5 Bundler respectively. In our thesis we mainly concentrate on d5DCL, d5designer and d5Bundler software.

This software is used to create the graphical user interface for the program logic that has been developed in the Code Composer Studio. Therefore it is necessary for the crane logic to have this interface, as the main goal of our thesis is to replace the dSpace system with the dasa control systems. We describe about each developing tools clearly in the chapter 5.

2.3 Automatic Conversion Tools:

The tools needed for the automatic conversion of Matlab code to the C code are to be investigated. These

tools must also be investigated in terms of efficiency and less time consuming procedures in the process of conversion. Agility MCS synthesizer is one among those tools that is investigated in specified terms. This automatically converts the Matlab files to the Desired C code.

Agility MCS synthesizer can be known in detail in the later sections of the report. The Matlab itself can also be a tool itself that can be used in the conversion procedure but is proved to be less efficient with student version software.

But, we have not done enough work in working with these tools in our thesis. We now continue our discussion with the crane kinematics and the geometry that would be the basic building blocks of the project and implementation of the algorithm.

(12)

2.4.1 Crane Geometry:

Lift arm Elbow arm 𝜃3 𝜃2 𝑑4 z Swivel arm 𝑂4 𝜃1 r

Figure 1: Forward crane geometry in two dimensions.

In Figure 1 the forwarder crane is represented in two dimensions. The swivel joint 𝜃1 is left out in our case in order to be able continue our presentation into the two dimensional case. The crane tip position, 𝑂4, can be represented in Cartesian coordinates, r and z or in terms of crane coordinates 𝜃2, 𝜃3 and 𝑑4from the Figure 1. From the crane geometry we have 𝜃1, 𝜃2and 𝜃3 are the angles of rotation at the links. The value 𝜃2, 𝜃3 reprensent various links such as swivel arm ,lift arm and elbow arm respectively. Thus the value of angles of rotation can be the parameters of choice.

Where, 𝜃1 is the swivel arm angle in radians, 𝜃2 is the lift arm angle in radians, 𝜃3 is the elbow arm angle in radians,

(13)

2.4.2 Crane Coordinates: 𝒚𝟐 𝒙𝟑 𝒙𝟐 𝑦1 𝜃3 𝑧3 𝑥1 𝑑4 𝜃2𝜃1 𝑧1 𝜃2 z 𝑂4 𝜃1 𝑧0 𝑥0 r

Figure 2: Crane coordinate according to Denavite-Hartenberg Table.

2.4.3 Forwarder Kinematics:

The forwarder kinematics can be derived using the methods presented in [9]. According to

Denavite-Hartenberg the transformation matrix T, transforming the coordinates between two subsequent coordinate system, from system i to system i-1, is given by the following relation as,

𝑇

_𝑖−1=

𝑇

_𝑖𝑖−1

=

𝑐𝑜𝑠𝜃

_𝑖

−𝑠𝑖𝑛𝜃

_𝑖

𝑐𝑜𝑠𝛼

_𝑖

𝑠𝑖𝑛𝜃

_𝑖

𝑐𝑜𝑠𝜃

_𝑖

𝑐𝑜𝑠𝛼

_𝑖

𝑠𝑖𝑛𝜃

_𝑖

𝑠𝑖𝑛𝛼

_𝑖

𝑎

_𝑖

𝑐𝑜𝑠𝜃

_𝑖

−𝑐𝑜𝑠𝜃

_𝑖

𝑠𝑖𝑛𝛼

_𝑖

𝑎

_𝑖

𝑠𝑖𝑛𝜃

_𝑖

0 𝑠𝑖𝑛𝛼

_𝑖

0

0 𝑐𝑜𝑠𝛼

_𝑖

𝑑

_𝑖

0

1

………. (1)

Denavite Hartenberg uses four parameters in the transformation matrix. The coordinate systems for the

(14)

The corresponding D-H table is given in Table 1.

DH-table, Crane 𝜽 𝒅 𝒂 Swivel arm, link1 𝜃1= 0 𝑑1= 𝑑1∗ 𝑎1= 0 ∝1=

𝜋 2

Lift arm, link2 𝜃2= 0 𝑑2= 0 𝑎2= 𝑎2∗ ∝2= 0 Elbow arm, link3 𝜃3= 0 𝑑3= 0 𝑎3= 0 ∝3=

𝜋 2

Extension link, link4 𝜃4= 0 𝑑4= 𝑑4∗ 𝑎4= 0 ∝4= 0

Table 1: Denavite-Hartenberg table for the crane considered.

Each of the four parameters is used in a simple transformation matrix in order to generate the eq(1). The simple transformation matrices are given by relation as,

𝑇

0

=

1

0 0 1

0 𝑎

0

0 0 0

0 0

1

0 0 1

𝑇

1

=

𝑐𝑜𝑠𝜃

−𝑠𝑖𝑛𝜃

𝑠𝑖𝑛𝜃

𝑐𝑜𝑠𝜃

0 0

0

0 1 0

0 1

𝑇

2

=

1

0 0 cos⁡

(𝑎𝑙𝑝ℎ𝑎)

0

0 cos⁡

(𝑎𝑙𝑝ℎ𝑎)

0 0 sin⁡

(𝑎𝑙𝑝ℎ𝑎)

0

0 cos⁡

(𝑎𝑙𝑝ℎ𝑎) 0

0

1

and,

𝑇

4

=

1

0 0 1

0 0

1 𝑑

0

1

(15)

The transformation from the 𝑂4 coordinate system to the 𝑂0 coordinate system is given by the relation as,

𝑇

40

= 𝑇

0

𝑇

1

𝑇

2

𝑇

3 ..………….……… (2)

𝑇

40

=

𝑐𝑜𝑠( 𝜃

2

+ 𝜃

3

)

0 sin⁡

( 𝜃

2

+ 𝜃

3)

−1

sin⁡

( 𝜃

2

+ 𝜃

3)

𝑎

2

𝑐𝑜𝑠𝜃

2

+ 𝑑

4

sin⁡

(𝜃

2

+ 𝜃

3

)

𝑐𝑜𝑠( 𝜃

2

+ 𝜃

3

) 𝑎

2

𝑠𝑖𝑛𝜃

2

− 𝑑

4

cos⁡

(𝜃

2

+ 𝜃

3

)

sin⁡

( 𝜃

2

+ 𝜃

3)

0

0 −𝑐𝑜𝑠( 𝜃

2

+ 𝜃

3

) 𝑑

1

+ 𝑎

2

𝑠𝑖𝑛𝜃

2

− 𝑑

4

cos⁡

(𝜃

2

+ 𝜃

3

)

0

1

The 𝑂4 i.e. the origin in the fourth coordinate system, is given in the (𝑂0, 𝑥0. 𝑦0. 𝑧0) system by the relation as,

𝑝

0

= 𝑇

0

𝑇

1

𝑇

2

𝑇

3

𝑂

4

The forwarder kinematics for the crane when 𝜃1= 0 is given by the relation as,

𝑝

0

=

𝑥

0

𝑧

0

=

𝑎

2

𝑐𝑜𝑠𝜃

2

+ 𝑑

4

sin⁡

(𝜃

2

+ 𝜃

3

)

𝑑

1

+ 𝑎

2

𝑠𝑖𝑛𝜃

2

− 𝑑

4

cos⁡

(𝜃

2

+ 𝜃

3

)

...(3)

This forwarder kinematics, when the swivel angle theta1=0, is a mapping from the three crane coordinates, theta2, theta3, and d4 onto the Cartesian coordinates x0 and z0.

The mapping clearly shows the non-linear function from the crane coordinates to Cartesian coordinates. The mapping will be used in a later section, but first the today’s way of crane control is introduced.

2.4.4 Today‟s forwarder operators:

(16)

(17)

3. PRISMATIC JOINT ALGORITHM IMPLEMENTATION

In the field of study Laute and In den Kleef has recorded the operators way of controlling the crane while loading and unloading logs, see [8]. In figure 3 and 4 the correlation between the extension boom, expressed by the crane coordinate d4, and the first boom and the outer boom, expressed by the coordinates theta2 and theta3, are shown. The correlation achieved for one operator is shown in Figure 3 and for another operator in figure.

Figure 3: Measurement of links velocities for a working forwarder for operator 1, from report by Laute and In den Kleef.

Figure 4: Measurement : Measurement of links velocities for a working forwarder for operator 2, from

(18)

In the measurement of the angular velocities it can be seen that the velocities of the extension boom are highly correlated with the first boom and the outer boom angular velocities, e.g., positive velocity of the extension boom is correlated with negative velocity of the first boom.

The suggested automatic system uses the found correlation in order to partly automate the crane control. In order to develop the prismatic joint algorithm we need to consider some intermediate functions in achieving the goal. Those intermediate functions include Jacobian Matrix, The gain parameter, the weighting function and the angular velocities of the various links which are discussed in the following sections

3.1 The Jacobian Matrix:

The Jacobian for the crane operating in two dimensional space spanned by (r, 𝑧 ), can be obtained by taking the derivative of the corresponding forwarder kinematics of the crane, eq(3).

Cartesian velocity equals the Jacobian times joint velocity. The derivation is performed when swivel joint is considered to 𝜃1= 0 and the coordinates 𝑥0 coincides with the coordinate r. The Cartesian velocity relation can be explained in the following relation as,

𝑣 = 𝐽 𝑞 𝑞

……… (4) Where

𝑣

is the Cartesian velocity and

𝑞

matrix representing the angular velocity of the each links.

Where

𝑞 𝑖𝑠

𝑞 =

𝜃

2

𝜃

3

𝑑

4 .

By taking the derivative of the left hand side and the right hand side of (3) the Jacobian cab be derived. The Jacobian Matrix is given by the relation as,

J(q)

=

−𝑎

2

𝑠𝑖𝑛𝜃

2

+ 𝑑

4

cos⁡

(𝜃

2

+ 𝜃

3

) 𝑑

4

cos⁡

(𝜃

2

+ 𝜃

3

)

sin⁡

(𝜃

2

+ 𝜃

3

)

𝑎

2

𝑐𝑜𝑠𝜃

2

+ 𝑑

4

sin⁡

(𝜃

2

+ 𝜃

3

)

𝑑

4

sin⁡

(𝜃

2

+ 𝜃

3

)

−cos⁡

(𝜃

2

+ 𝜃

3

)

...(5)

We now proceed in discussing to the other functions that are involved in building the algorithm of the crane.

The Cartesian velocities in two dimensions case is given by the relation as follows,

(19)

This can be calculated as follows,

𝑣 = 𝐽 𝑞 𝑞 Where 𝑞 = 𝜃 2 𝜃 3 𝑑 4 ………..…..……….……..(7)

The values of the angular velocities can be obtained from the sensors. In the development of the algorithm the maximum joint velocities are also considered𝜽 𝟐𝒎, 𝜽 𝟑𝒎, 𝒅 𝟐𝒎.

3.2 The Weighting Function and Gain parameter:

A weighting function is introduced in order to take into account the prismatic link position and velocity. The weighting function is given by,

∧ 𝑑4, 𝑑 4 = 𝑑_4𝑒−𝑑₄ 𝑑4𝑒−𝑑4𝑐 , 𝑑4> 𝑑4𝑐 & 𝑑 4> 0 1 , 𝑑4> 𝑑4𝑐 & 𝑑 4< 0 1 , 𝑑4< 𝑑4𝑐 & 𝑑 4> 0 1 − 𝑑4𝑒−𝑑4 𝑑4𝑒−𝑑40, 𝑑4> 𝑑4𝑐 & 𝑑 4> 0 ………(8)

Where 𝒅𝟒𝒄the centre is position of the prismatic joint and 𝒅𝟒𝒆 is the end position of the prismatic joint Thus these equations are a part of the algorithm for the desired prismatic link that should be implemented in the dspace system previously. The algorithm should to be implemented in the Code Composer Studio environment. This software here is known as the software 1A.

3.3 Parameters Chosen for the Algorithm Implementation:

The operator selects the parameters of his choice in order to operate the crane. The operator is provided

with a display screen where the values can be configurable. Firstly, we need to deduce which parameters we need to consider that are needed to be modified. Care should be taken in choosing the parameters and understand the significance

(20)

Then the gain parameter 𝒌 for the parameter of choice is also considered. The value of 𝑘 is deduced from the desired prismatic joint velocity. The parameters must be made configurable and the programming is developed according to the parameters of our choice to reach the desired goals of the automatic extension link.

Algorithm associated parameters

Description of Parameter

𝒅

_𝟒

Joint Velocity of the Extension link.

𝒅

_𝟒𝒅

The Desired Prismatic Joint Velocity. 𝐝𝟒𝐜 The Centre Position of the Prismatic Joint.

𝒅𝒅𝟒𝒅𝒆𝒔𝒊𝒓𝒆 The desired values of the prismatic link velocities.

𝐤 The scalar gain parameter constant

𝐚𝟐 The length of the lift arm link of the crane in meters

𝐝𝟒 Length of the Prismatic Joint

(21)

(22)

4. RK-62 PARAMETERS

Figure 5: The RK-62 parameters which we need to be implemented in the Crane lab set up.

The algorithm implemented is now tested in the lab environment. The RK-62 parameters are the real crane parameters where in which we set the algorithm implemented for the actual physical crane. The elbow arm, lift arm and the extension link are chosen from the given data sheet in order to operate the crane. The gain parameter and the lift arm are the two parameters of choice in our algorithm. These parameters could be explicitly varied on the operator’s choice. When we deal with rk-62 parameters, we may have more numbers of parameters in picture. Thus the choosing the parameters could be a crucial thing in the implementation in the real crane scenario.

The values in the dasa lab testing carried out which will be in the following sections must be modified with the RK-62 parameters so that we need to make our algorithm run with the rottne crane available at the crane lab. From the Figure 5 we can deduce the RK-62 parameters those to be implemented in the control algorithm. The parameter specifications from the RK-62 crane can be stated as the gain parameter, the length of the

(23)

(24)

5. DASA LAB TESTING ENVIRONMENT

The dasa lab testing of the algorithm is carried out using the Visual Studio C++ software initially. We need to verify the simulations of the algorithm of the automatic extension of the prismatic link in the dasa environment. With these valid simulations we need to proceed to the crane lab for the practical crane issues of operating the crane links.

The algorithm is written in this environment using all the functions and calculations available in the algorithm. The algorithm written needs to be verified for the errors and the errors needed to be resolved. When the algorithm is made to compile and the errors are eliminated, and then the process of debugging is initialized. The whole code is debugged and the results at each output are verified. The debugging process also gives us some errors and they need to be resolved and we should continue this process till we obtain desired results. Once sure of the results obtained after the debugging process are valid, the development of desired algorithm for running the crane of operator’s choice is ready for the simulation.

This code is now sent into the Code Composer Studio (CCS) for making it DSP processor compatible which is the main focus of our project. Then needs to check the algorithm for errors and make it errors free even in Code Composer Studio environment.

(25)

5.1 d5DCL:

It is a programming tool for establishing application linked DCL code. We configure descriptions of component parts, I/O specifications, variable and control logic.

Figure 6: The d5DCL window of the logic development for the crane link parameters. 5.2 d5Designer:

(26)

Figure 7: The d5Designer window which provides the graphical user interface. 5.3 d5Bundler:

It is the software that collects MMI, DCL packages and an access file and sender to dasa5 main controller. Those files which are collected are sent into a distribution file which serves as the gateway communication between d5developer and the dasa5 microcontroller.

Figure 8: The d5bundler which bundles the various d5 software and links to the dasa equipment.

5.4 dasa5 System Units:

The physical units of the dasa equipment comprises of different I/O units, Main Controller and base units those can be seen in detail in the following chapters.

5.4.1 d5CI12 Unit:

This is capable of doing time critical mathematics in real time scenario. These may include as bucking

(27)

Figure 9: The Physical structure of d5CI12 unit. 5.4.2 TFT Monitor:

It is a mobile windows-based display monitor developed for the suitable environment of heavy vehicle.

These units are now used for communicating with each other in the physical environment in the real crane

scenario.

Figure 10: The monitor screen display at the dasa lab for verifying the simulations. 5.4.3 d5MC (Main Controller) Unit:

It comprises of a DSP processor for controlling the remaining units through a CAN bus to the machine,

(28)

Figure 11: The main controller unit d5MC for controlling the signal flow. 5.4.4 d5IO12 Unit:

I/O units which contain 12 independent programmable inputs or outputs, of the inputs are current, voltage,

frequency, counters and outputs as PWM for servo values, digital & analog and with Max rating 3A.

Figure 12: The d5IO12 input-output unit in the dasa equipment.

5.4.5 d5IO48 Unit:

(29)

Figure 13: The input-output unit d5IO48 of the dasa equipment.

5.4.6 Sensors:

(30)

(31)

6. THE CRANE LAB SETUP ENVIRONMENT

6.1 Lab Crane Set Up With the Dasa equipment:

The physical crane setup is kept at the Vaxjo University which is provided by the ROTTNE INDUSTRII

AB. We need to replace the existing dSpace system with the dasa5 equipment for running the crane. The modules such as d5MC, d5IO48, d5IO12 and d5CI12 units are needed along with CAN buses do the Dasa equipment. The software which we develop is called to be Crane Box that is needed to be en suited in the d5CI12 base. We need to control the crane by using joysticks according to the operator’s choice. The different parameters needed to be configurable according to the operator choice, are needed to be configured by the d5 equipment. There are also some important units such as mounting of sensors at the setup and assignment of proper ID tags for the proper signaling of the systems is done with utmost care. Therefore we shall see in detail are the interconnections in the crane lab setup in the following explanation.

The Crane Box unit is interfaced with the three sensors Theta2, Theta3 and d4 with the sref, S90, S00 connections of the Port A. The Port B gets the signals from the d5 MC unit and an ID Tag. The d5 MC unit is interfaced with the Crane Box with the CAN buses of 120 Ohm resistance. The d5 MC unit is also interconnected to the d5IO48 unit with the PWR/IO. The d5IO48 unit is also connected to an ID Tag as well as to the Crane Box unit with the 120 Ohm resistance. The d5IO48 unit is interfaced with the left joystick of the left Dsub female pin through the Port A. The port B is interconnected with the right joystick through the right Dsub female pin. Thus by using these joysticks the operator controls the crane of his choice.

The d5IO12 unit is interfaced with the main controller unit through the CAN bus which is specifically CAN1 Bus. This in turn connected to the crane valves which comprises of 8 current outputs to the crane set up. The sensors mounted on the crane were connected at various links and we get the desired outputs feedback from them. The interconnections can be made with at most importance because of the previous dspace system connections. We also should take at most care about the hydraulic pressure pump while we are setting it open and closing the valves while operating the crane.

Key Aspects,

 Mounting the sensors.

 Switching the dSpace to Dasa equipment.  Handling the pressure pump.

(32)

6.1.1Physical connections at the crane lab:

+

yellow +Red +yellow +Red +Blue

+brown

+

Green +Brown +Green sref {8}

sref

PWR 24v

S00 s90 GND

GND

PWR 5V

+4 +1 +12 +2 +11 +3 tags Idtag=1

+5

+7 CAN1 bus

+8

+

9 +4 +12 CAN2L {0,5 m} +13

+1

CAN2H {0, 5 m} +5

+11

+2 PWR2{0,5 m}

CAN2L {0,5m} CAN2H{0,5m} GND{0,5 m} +9 PWR2{0,5 m} +12

+3 GND{0,5 m}

+ 11 +13 GND

+2 X +1

+cyan +black +12 Y +6

+3 Z +5

+20 +5V +3

Figure 15: Overview of the lab crane system with dasa5 system

Crane Box

Port A

Port B

d5MC

CAN

PWR/IO

120 ohm

IO48

CAN/PWR

120 ohm

Port B

left: Dsub female Left: Joystick Same pinning as for port A

Port A

Right: Dsub

female Right: Joystick

3: IdTag

2: Id

Tag

d5IO12

Valves (8pcs)

8Current Out

2mcables

(connector)

(33)

SL NO Notation in Physical Connections

Description of each Notation

1 Sref The zero position signal line for the sensor of 24 Volts.

2 S90 The signal but is delayed by 90 degrees for the sensor of 24 Volts.

3 S00 The signal line for the sensor of 24 Volts.

4 GND The ground for the power supply.

5 PWR The power rating for the sensor .

6 Id Tag Provide a unique ID for the d5IO12 UNIT.

7 2:Id Tag Provide a unique ID for the CRANE BOX.

8 3:Id Tag Provide a unique ID for the IO48 UNIT.

9 Right Dsub The Joystick controller signal at Port A of d5IO48.

10 Left Dsub The Joystick controller signal at Port B of d5IO48.

11 CAN 1 The bus structure for data transfer at d5IO12 and d5MC.

12 CAN 2L The CAN 2 LOW.

13 CAN 2H The CAN 2 HIGH.

(34)

6.2 Using the dSpace System in Crane Lab Set up:

Crane

Theta2, theta3, d4… 8 current outputs

{8}

(35)

In the lab crane system, various units are interfaced with the actual crane. Mainly a dSpace system where in which the crane controls software is implemented previously. The setup include a d5MC (main controller), d5I012 base, crane values, sensors, joysticks for the operator.

The setup is the physical implementation where the dSpace systems interface with firstly the control signal from the joysticks of the user end and secondly the sensors through the parameter of the choice are operated automatically and thirdly with the dasa software’s.

At the lab provided with a crane which is of the ROTTNE INDUSTRII AB Company. In the Lab environment the crane, is mounted on a well settled base which is in equilibrium position. The crane set up is provided with the hydraulic pressure pump generator through which the necessary oil pressure is sent to the crane. The crane is provided with the necessary electricity and is plugged in to the power supply.

The crane set up is provided with the dasa5 system units, d5 MC which is the main controller which is equipped with the d5 software which provides the user interface with the crane setup. This has the micro controllers and the micro processors for the communication with the other devices. The d5MC unit is connected to the d5CI12 base through the CAN buses where in which the input and output operations are carried out. The signaling and controlling operations are generated to the dSpace system, the joysticks and to the sensors. The dSpace system is connected to the d5CI12 base in which the actual software is executed for the crane control. This dSpace system is made to function with the SIMULINK and MATLAB programming formerly. This is evident that now we need to replace the dSpace system with the new d5 system equipment in which the Crane Control SW is implemented through the C programming for the crane operation.

The dSpace system is further connected to the sensors and also to the joysticks for the operator. Firstly the joysticks can be controlled by the operators which are assigned with the parameters of our choice that are needed to be controlled on the crane.

Secondly the sensors are connected at various links through which we obtain the feedback signals for each movement in the respective links of the crane. In specific we are having three links and we can have only 2 angles and a length, such as theta 2, theta3 and d4 respectively. Finally the crane valves which are connected to the d5CI12 base, through which provided with 8 current output vales of the crane setup in order to control the crane.

(36)

6.3 Replacement of dSpace System with dasa5 Equipment:

.

Theta2, Theta3, d4…… 8currentout {8}

CAN bus

112

CAN bus

Figure 17: Overview of the lab crane system with dasa5 system

(37)

The Dasa5 implementation is carried out with enough modifications and replacements of the various units which are present already in the crane lab setup with the dspace system quipment.As we are already sure that we need to replace the dspace system with the dasa5 equipment in operating the crane is the foremost task. This d5 implementation environment comprises of various units which include as, d5IO48 unit,d5MC unit,Crane Box in which the SOFTWARE 1A is implemented,d5CI12 unit,joysticks, the sensors,the CAN buses and the crane valves.These units are interfaced with slight modifications in the interconnections and are implemented at the lab crane environment.

The joysticks are connected to the d5IO48 unit which is an I/O unit generally use dfor low power applications.The unit is in turn connected to the d5 MC the main controller unit which comprises ofo the DSP processor for controlling and signalling various other units in the environment and is in turn interfaced with the crane control sofware, device drivers and the interface of the dasa5 equipment.The d5MC unit is further connected to the d5CI12 base which is also an I/O unit to control varoius parameters restricted within the crane set up. The sensors are connected similarly as in the crane lab set up, where they are connected at various links.The fuction of the sensors is also similar, as they send the feedback signals with some specifuc valuesto

the operator.The crane valves are connected to the d5IO12 base where we obtain 8 current outputs. The sensors specifications in the lab crane set up are different to the D5 implementation environment.In

the case of the lab crane system we are using 5 Volts sensors whereas in the case of the dasa5 implementation we have the 24 Volts specfication sensors.This helps us to know that the dasa equipment are spacifically designed for that particular voltages only.But when we compare the dspace system in the lab crane set up we are using quite low specification sensors I the operating the crane. The dasa5 implementation is carried out along with the SOFTWARE 1A for the controlling of the crane accordiing to the operators choice. These are all the necessary modifications done to the crane set up in order to test the crane functioning.There should be also precaution measures to be taken while operating the crane such as the situations of hitting the ground and the the side way movements.

6.4 Typical crane functionality:

(38)

Figure 18: Overview of the sensor voltage at dd4 display on dasa5 system software

In the Figure 18 the sensor voltage at the dd4 displayed for the each change in the angles as well as the change in the extension link when they are switched back and forth. The minimum voltage and the maximum voltages are set at 56000 to -56000 according to the figure above. This voltage could be changed using the DCL configuration settings by assigning the required voltage settings. We also need to know the sensors specifications such that they may be supporting to their maximum capacity only. When we go on to operate the crane we can obtain the voltage values in mille volts on the screen and accordingly we can perform our tests on the algorithm we excecute.

(39)

Figure 19: Overview of the lab crane system sensor voltage at d4 display in dasa5 system software

In the Figure 20 the DCL software is discussed in configuring the voltages and currents of the dd4 and d4 displays. The window shown above allows configuring the parameters and the respective values to them. We know can have the brief overview of configuring the parameters. In the window shown we have the main menu which includes the file menu, edit menu, view menu, project menu, tools menu and help. We create the new project where in which we configure the system variables and the respective assignments of the addresses. Once when we are sure of all the system variables are correctly assigned we then scale them as the next step. We scale the respective parameters click on each system variable and check out for each of the cell information and then we click on configure scale.

(40)

Figure 20: Configuring the sensor voltages in dasa5 software at the crane lab environment

. The functionality of the crane can be viewed both in the theoretical basis and the practical basis. The sensor voltage also plays a major role in filtering out the values and displays the stable values. The 10 ms sampling time is another major concern in taking the value for that amount of time.

(41)

(42)

7. AGILITY MCS SYNTHESIZER TOOL

Agility MCS is an automatic generator of ANSI C code from MATLAB source. This tool was designed by

Agility Design Solutions which are formerly known as CATALYTIC INC. The CATALYTIC Inc. combined with CELOXICA ESL and formed a new organization named to Agility Design Solutions. It also develops C-code and FPGA-code synthesis. Its mission is to lessen the time to implement, develop and validate signal processing algorithms. Therefore agility MCS became user friendly due to its time saving approach to complex signaling in signal processing applications. It supports different toolboxes such as MATLAB communications toolbox, MATLAB signal processing toolbox and MATLAB image processing tool box. It runs in Microsoft Windows (2000 or XP) and 32-bit or 64-bit Linux (Red Hat Enterprises version 3 and 4).it requires the product such as MATLAB(R 12.1 or later). Thus Agility MCS runs successfully once we create this entire environment and gives us the optimized results. We deal in detail the procedure of conversion along with some peculiar features and characteristics of Agility MCS software tool in the further sections.

7.1 Creation of C code automatically:

Agility MCS helps MATLAB users to get required C code templates of the M-code models respectively. There is some help to the user of MATLAB in the implementation of the algorithm. The Agility MCS takes off once site from manually writing C-code which proves to be tedious, time consuming job and vulnerable to errors. It creates a platform named as MATLAB test bench in common for repeated algorithm validation. It develops algorithm models that flexibly cope up with already present flows and performs on any ANSI C-code platform. Beside this the agility MCS defines a specific white-box algorithm hand-off for signal processing applications. Agility MCS family is an automated process which helps the programmers to start even through the algorithm is modifying current time. Agility MCS also prepares prototypes and development in the algorithm within minutes rather than in days or week. Thus, these exciting and unique features are attracting many programmers in MATLAB to use agility MCS synthesizer for many signal processing and communication applications.

7.2 Agility MCS Characteristics:

Generally at the time of conversion of the desired M-code to the C-code using the Agility MCS, the

obtained C-code exhibits some unique features in the code hierarchy. Thus proven to be user friendly and has become an easier choice for the programmers to work with. Therefore we shall look into those features in detail in the following.

(43)

 It does change the function names and also the variable names present in the M-code that is subjected to conversion.

 When the function hierarchy and file stricter are compared from the M-code it remains similar and helps the programmers to validate the code.

 Besides conversion it preserves the M-code as comments in the obtained C-code.  It includes the pre-implemented C-code.

 Agility MCS through graphical user interface its cross validate MATLAB code and the C-code obtained.

7.4 Agility MCS Synthesizer Block Diagram:

Target Custom Processor Algorithm prototype

ESL

System Simulation Verification

Software Application

Figure 21: Agility MCS block representation of the synthesizer.

M-code

Agility MCS

(44)

(45)

8. CONCLUSION

The aim with this thesis project was to implement an automatic algorithm for a forwarder crane. The

algorithm is implemented into a Dasa control hardware called Crane box including interface routines for communication with the D5 box. The implementation was performed manually. The functionality is partly tested by use of three methods.

First is the software tested in the crane laboratory at Linneus University in Vaxjo. The implementation has been tested by the skilled crane operators from the Rottne Industry AB.

Secondly the software is tested by interaction with the simulated crane, also implemented in the Crane Box system.

Thirdly the software output for some examples has been compared to the dSpace implementation of the algorithm.

(46)

9. REFRENCES

[1] Westerberg S., Manchester I.R., Hera P.L.,and Shiriaev A. Virtual environment teleportation of a hydraulic forestry crane. In International conference on Robotics and Automation, Pasadora, C.A, USA, May 19-23, IEEE, 2008.

[2] Lofgren B., “Kinematic Control of Redundent Kunckle Booms eith Automatic Path-following Functions”, Doctrol Thesis, KTH, 2009.

[3] Sigvardsson M. and Olsson T., “Modelling and Simulation of a Hydraulic Crane”, Master thesis, University of Kalmar, February 2005.

[4] Ekevid T. „On optimal control of hydraulic cranes”, Proc. 19th Nordic Seminar on Computational Mechanics, NSCM-19, Lund, Sweden, October 2006, pp 107-110.

[5] Fazululla M. and Srikanth K., “Mathematical Modeling and Simulation in Dymola of a laboratory crane”, Master thesis, University of Kalmar, June 2006.

[6] Heinze A.,” Modelling and Simulating of Laboration Crane, friction and dynamics”, Master Thesis Report, Växjö University, 2008.

.[7] Zhamykhanova A. B. “Modelling of a laboratory crane”, Master thesis, University of Kalmar, 2008. [8] Laute M, and In den Kleef P,(2010), Measurement of the link operation of a working forwarder, Project report, Vaxjo University,{to appear 2010}.

[9] Spong M.W., Hutchinson S., and Vidhyasagar M., Robot Modelling and Control, John Wiley and Sons,2006.

(47)

(48)

GRAPHICAL RELATION

Graphical diagram:

Represents relation between the C-Code and Algorithm.

Figure 22: Overview of the relationship between the algorithm and the code.

(49)

The desired prismatic link velocity can be obtained from the functions of the crane geometry.Then we formulate the Jacobian matrix by taking the derivative of the forwarder kinematics obtained from the transformation matrices. This could be related to the code is qdot2V function. In this function not only evaluates the Jacobian matrix but in tern the Cartesian velocities of the links of the crane.

The angular velocities can be obtained from the sensors of the crane. Thus we need to consider the

maximum angular velocities of the links of the crane. These also could be included in the qdot2V function.

These are all the functions that were included in the algorithm.

The gain parameter 𝐤 is also included in the qdot2V function. The gain parameter is a scalar quantity that could be multiplied to the Jacobian matrix in this function. Then the crane tip in order not to hit the end positions we develop a weighting function that could be included in the algorithm.

The Weighting Function is related to the code as the Lambdafun function for the algorithm implementation. These were used to compute the dd4desire values of each crane experiments that were performed at the dasa laboratory. These could be related to Crane Experiments.c in the code. The

dd4deridedfunc is rela.ted to the dd4desire values evaluation of the algorithm.

Then using the dasa software we need to develop the user interface for the links of the crane in functioning of the crane. These were all used in order to compute the 𝑑 4𝑑 velocities of the crane. The User.c can be used to configure the variables for the crane laboratory and also provides the graphical user interface for the display screen.

This User.c of the code is related to the user interface of the algorithm. Thus the extension boom velocities are calculated in the algorithm with all these supporting files in the code development. Thus we have brought the relation between the algorithm and the C code that would help in understanding the UML diagram which is discussed in the Figure 19 above.

(50)

(51)

10. RESULTS

Figure 23: Dasa lab testing environment simulations with crane lab experiments.

(52)

Figure 24: Crane lab testing environment output curve from the d5 Analyzer.

This was the output curve that was taken from the d5 Analyzer and we can notice a smooth curve when the same testing of picking of a log from the ground is performed. The curve is a smooth decreasing one that could be compared against the matlab or the lab simulations as we said before. We have to perform different steps while testing the crane functionality. We have taken out the manual steps in the figuring out of the outputs from the crane functionality. We have taken up the process of identifying the zero positions of the links and then we started to read the inputs and respectively we have drawn the outputs. We have figured out various values of the thea2, theta 3 and the respective dd4desire values at that point of time. This all has bee possible only through the usage of d5 Analyzer.

(53)

Theta 2 angle(radians) Theta 3 angle(radians) dd4 desire velocity(m/sec) 83 -80 1168 80 -78 8132 77 -73 7441 75 -69 6014 71 -64 4473 60 -61 4813 53 -54 4519 46 -49 3627 42 -45 3133 28 -33 2416 11 -15 2011 6 -5 1218 3 -1 0826 1 3 0413

Table 4: The algorithm dd4dersire values that were taken from the crane are illustrated.

(54)

(55)

Blekinge Tekniska Högskola SE–371 79 Karlskrona Tel.vx 0455-38 50 00 Fax 0455-38 50 57

11. UML DIAGRAMS

ANALYSIS DAIAGRAM:

Figure 25: Analysis diagram gives the relation between the physical and coding parts.

act algorithm

EA 7.5 Unregistered Trial Version EA 7.5 Unregistered Trial Version EA 7.5 Unregistered Trial Version EA 7.5 Unregistered Trial Version EA 7.5 Unregistered Trial Version EA 7.5 Unregistered Trial Version

ActivityInitial

Calculate deriv ativ es theta2 theta3 UpdateInputValues AndDependencies () multiplication deriv ativ es(dtheta2,dtheta3,dd4)

multiplication d4 Jacobian Matrix qdot4m matrix qdotm matrix multiplication Transpose Matrix V4m matrix V23m matrix V23 matrix Normalization of V4m matrix normalization of V23m matrix

div ide both matrices

(56)

Blekinge Tekniska Högskola

SE–371 79 Karlskrona Tel.vx 0455-38 50 00 Fax 0455-38 50 57 COMPONENT DIAGRAM

Figure 26: Component diagram gives the relation between the physical and coding parts. cmp Components

DSP IMPLEMENTATION OF A-CONTROL ALGORITHM FOR A FORWARDER CRANE