Content uploaded by Claus G Sørensen
Author content
All content in this area was uploaded by Claus G Sørensen
Content may be subject to copyright.
1
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant
Nursing
R.N. Jørgensen1, C.G. Sørensen1, J. Maagaard2, I. Havn2, K. Jensen1,
H.T. Søgaard1, and L.B. Sørensen2
1Aarhus University, Institute of Agricultural Engineering. Department of Agricultural
Engineering. Research Centre Bygholm, Schüttesvej 17, DK-8700 Horsens, Denmark.
2Vitus Bering, University College, Chr. M. Østergårdsvej 4 C, DK - 8700 Horsens, Denmark.
E-mail: Rasmus.Joergensen@agrsci.dk
ABSTRACT
Danish organic outdoor gardeners today use 50-300 hours per hectare for manual weeding.
Through automatic controlling of an existing commercial machine this often heavy and cost-
consuming weeding will be eliminated. At the same time, a fully-automatic registration of field
activities will contribute to the efficient implementation of EU directive 178/2002 concerning
traceability in the primary production and thereby enhance the food-safety in the production
chain. A radio controlled slope mower is equipped with a new robotic accessory kit. This
transforms it into a tool carrier (HortiBot) for high-tech plant nursing for e.g. organic grown
vegetables. The HortiBot is capable of passing over several parcels with visible rows
autonomously based on a new commercial row detection system from Eco-Dan a/s, Denmark.
This paper presents the solutions chosen for the HortiBot with regard to hardware, mechanical-
electrical interfaces and software. Further, the principles from a Quality Function Deployment
(QFD) analysis was used to carry out the solicitation, evaluation and selection of most qualified
design parameters and specifications attained to a horticultural robotic tool carrier. The QFD
analysis provided a specific measure to evaluate each selected parameter in terms of satisfying
user requirements and operational performance aspects. Based on a combination of importance
rating and competitive priority ratings important user requirements include easy adaptation to
field conditions in terms of row distance and parcel size, profitability, minimum crop damage
during operation, and reliability. Lesser importance was attributed to affection value, attractive
look, the possibility of out of season usage, and the use of renewable energy.
Keywords: Machine design, machine specifications, Quality Function Deployment (QFD),
robotics, tool carrier
1. INTRODUCTION
Within outdoor gardening, weeds are today a major problem, especially for early sown or
transplanted crops with a slow growth rate, like carrots and onions. Weed control can either be in
the form of mechanical inter row combined with intra row pesticide application (90% of the total
outdoor gardening area in Denmark) or mechanical inter row combined with manual intra row
weeding (10% of the total outdoor gardening area in Denmark). There is however, an increasing
demand from the consumers and the society to reduce the pesticide use in order to minimize the
impact on flora, fauna, aquatic system, and working environment.
2
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
Depending on the weed intensity, Danish outdoor gardeners use 50-300 h/ha for manual weeding
in onions and carrots (Ørum and Christensen, 2001; Melander and Rasmussen, 2001). This is
cost-intensive not only in direct labor costs but also in form of labor allocated to this one
operation relative to other urgent tasks within the growing season. Further, there are often
difficulties associated with procuring the necessary labor.
1.1 Robots within Plant Production and Outdoor Gardening for Weed Control Today
With regards to relevant weeding robots, worldwide, there exists only a few today. In Denmark,
there is a prototype called GreenTrac, which is designed as an environmentally sound tool carrier
for organic outdoor gardeners. Currently, the GreenTrac is not matured for production and is
unnecessary big for most tasks (Sørensen and Frederiksen, 2002). In Sweden, there is a robot for
intra row weeding in sugar beets (Åstrand and Baerveldt, 2002). Israel has a multi-functional
prototype robot for transplanting and spraying (Edan and Bechar, 1998). In England, an outdoor
gardening robot has been developed which is capable of passing over parcels of row crops (e.g.
Hague et al., 1997). However, it cannot perform proper field work.
1.2 Today’s Technological Barriers
The majority of agricultural prototype robots base its navigation on high precision satellite
position systems, and on field and crop maps. Hence, it is relatively information demanding and
complex to work with. From the operator viewpoint, commercialization of a field robot requires
that it is significantly simpler to operate (Callaghan et al., 1997; Jørgensen, 2005).
A technological development of weeding robots depends, apart from the market situation, to a
great extent on technological barriers and comparability with the existing technological stage.
Kassler (2001) lists barriers that have retarded the exploitation of computer-controlled machines
in agriculture, e.g.: Insufficiently robust mechanical technology; Costly mechanical technology;
Limited capability; Basic knowledge to create technology as dexterous or as skilful as that of a
trained worker is currently unavailable.
1.3 The Voice of the Customer
Within the concept of Total Quality Management (TQM), a number of tools have been adapted
to assist the process of customer driven planning and engineering for product development
(Cohen, 1995). One such tool is the Quality Function Deployment (QFD), which has as its
primary goal the translation of customer requirements into technical requirements of each stage
of the product design and production (Chan and Wu, 2001; Crowe and Cheng, 1996). The
process involves identifying customers’ requirements for a product (WHATs), customers’ view
on the relative importance of these requirements and the relative performance of the intended
product and the main competitors on these requirements. Also, the complete QFD process
includes translating the customer requirements into measurable engineering requirements
(HOWs) through careful evaluations performed by technicians recognizing the relationships
between customer requirements and engineering characteristics.
3
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
1.4 Aim and Deliverables
The consumer driven demands to reduce the pesticide usage increases the demand for
mechanical weed control as a way to avoid costly hand weeding. Within a few years, new
environmentally sound technologies are expected to replace hand weeding. However, this creates
a demand for a mechanical unit which will be able to carry future high precision weeding tools
with a low constant speed and high precision.
The aim of this paper is to describe a developed robust horticultural robot called HortiBot, which
will have the following main characteristics:
• Capable of passing over several parcels with visible rows autonomously based on a
commercial row detection system with no or minimal use of Global Positioning Systems
(GPS).
• Unskilled workers will be able to operate the basic functions of the HortiBot followed by
attending one hour of training.
• All operational data is automatically sent to an internet based database.
• The operation of the HortiBot is documented in terms of feasibility, operational capacity,
and economy.
Further, the objective is to identify qualified user requirements for the design of a robotic tool
carrier to be used carrying various implements for plant nursing.
1.5 Safety Emphasis
Through automatic regulation of an existing commercial machine, heavy and cost-consuming
weeding is eliminated. Further, fully-automatic controlling will contribute to the efficient
implementation of EU directive 178/2002 concerning traceability in the primary production and
thereby enhance the food-safety in the production chain.
2. MATERIALS AND METHODS
The project is coordinated by The Danish Institute of Agricultural Sciences, Department of
Agricultural Engineering, Denmark, with expertise within technologies for precision weeding,
robot technology for agricultural purposes, and machinery management.
The additional partners in the project are Vitus Bering, Denmark, with competences within
hydraulics, electrical control, and software development; Special Maskiner, Denmark, with many
years experience within specialized machinery to nurse green areas; Eco-Dan a/s, Denmark,
which is the leading supplier of vision based solutions for automatic tool guidance within row
crops; The horticultural enterprise Inge-Marienlund, Denmark, which is the largest producer in
Denmark of garden lettuce, china cabbage, and organic onions, grown as a part of the
approximately 170 ha farmed according to organic principles.
4
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
2.1 Hardware
The hardware design is modular and will to the largest extent be based on standard components,
making tailored components the last resort.
2.2 Software
The software is based on open source and open standard principles. Further, the developer’s kits
for the software environments should be easy available and inexpensive to acquire.
2.3 The Voice of the Customer
The overall QFD approach involves the ranking of technical specifications in relation to their
degree of contribution to the fulfillment of customer or user requirements. In other words, the
requirements of various interested parties are transformed into a description of the technical
specifications. Akao and Mazur (2003) defined QFD as "a method for developing a design
quality aimed at satisfying the consumer and then translating the consumer's demands into design
targets and major quality assurance points to be used throughout the production phase". The
analysis steps in this paper focus on: 1) determining customer requirements, 2) ranking the
requirements, and 3) competition benchmarking. For further details see Sørensen et al. (2006).
2.4 Selected Competitive Tool Carrier Systems for Weed Control
The HortiBot tool carrier – see also Jørgensen et al. (2006) - was compared with possible
competitive tool carriers, here the GreenTrac tool carrier (Sørensen and Frederiksen, 2002) and
the tractor equipped with Auto Guidance by AutoFarm – see table 1.
Table 1. Competitive tool carrier systems
HortiBot is a future commercial produced and
robust tool carrier. It will enable an automatic
execution of one-sided repetitive weeding for
outdoor gardening. The HortiBot will be able to
carry light weeding tools for parcels of 5–6 rows.
No prior planning is needed before starting a
weeding job, as the steering is primarily based on a
computer-vision-based guidance system. Typically,
the operator is an unskilled worker, whose primary
job is to monitor one or several weeding robots
instead of performing the labor-intensive work
manually.
GreenTrac is a future tool carrier to be used in the
growing season with light tools such as an inter-
row cultivator for row crops. Without human
assistance, it operates performing light work. The
5
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
GreenTrac navigates within the field by use of
high-precision satellite navigation, requiring that
the exact parcel positions and each crop row
position must be known beforehand. Each job is
planned at the office and then transferred to the
GreenTrac’s computer. For safety reasons some
sort of monitoring will be necessary. However,
several vehicles can easily be surveyed by the same
person.
The tractor is equipped with AutoFarm RTK
AutoSteer, which enables machine control for
repetitive treatments in the field with an accuracy of
3 cm. With this system, the parcels can be placed in
the same locations year after year, reducing the soil
compaction of the growth media. AutoFarm RTK
AutoSteer is easy to learn and to use for most
operators familiar with tractors. The job is planned
at the office on an ordinary computer and then
transferred to the tractor. A driver is required to
perform turns at the headlands and to control the
tools on the tractor.
3. RESULTS
The best suited platform identified as offset for a serial produced, reliable, and robust robot for
horticultural weeding was found in Spider ILD01. Spider ILD01 is a slope mower for
maintenance of uneven terrain with slopes up to 40° and is developed and produced by Dvořák
Machine Division, Czech Republic. The propulsion of the four wheels is driven by a central
hydraulic motor and the steering by a central electrical DC motor. The Spider is remotely
controlled by an operator and is changing its heading by turning all 4 wheels in parallel. Hence,
the orientation of the vehicle is not controllable, which will be a necessity for future operation
within row crops. This transformation of the conventional Spider slope mower into a tool
carrying and autonomous robot for horticulture is detailed in the following.
Visually, the changes to the original Spider ILD01 slope mower are minimal as a result of
transforming it into the HortiBot as illustrated in figure 1.
6
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
Figure 1. Illustration of the HortiBot equipped with individual wheel control, 3D row vision
system, and lift arms.
3.1 Hardware
Overall, the main change to the Spider slope mower has been transforming the joint wheel
control to individual controllable wheel modules. Each wheel module consist of a hydraulic
motor for propulsion, a DC motor for steering, speed and wheel angle sensor, and a control
module. The engine is also controlled by a control module, a lift arm with a control module is
mounted, and a central HortiBot Control Computer (HCC) has been mounted. The
communication between all units is based on a proprietary high speed CANbus. A joint control
module based on a 16 bit Atmel AVR microprocessor has been developed for the 4 wheel
modules, the engine control module, and the lift arm module. The overall mechanical setup and
electrical interfaces of the HortiBot can be seen in figure 2.
7
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
2 x camera
Manual controls
ISO bus
Digital signal processor
Actuator
Hydraulic on/off
Tacho sensor
Steering motor control
CAN bus
Cutter
Horn
Inclination alarm
Safety loop
LED display
Fuel valve
RS-232
RS-232
RS-232
Ignition
Engine temperatur
Hydraulic control
Engine RPM control
Engine module
Microcontroller
Hortibot
Control
Computer
Laser Range Finder
GPS Unit
Microcontroller
GPRS Modem Internet
Database
Connector Connector
LAN
Wheel module
Microcontroller
Lift module
Microcontroller
Vision module
Reciver module Transmitter
module
Figure 2. Hardware mechanical-electrical interfaces. The full lines indicate hardware units
mounted on the HortiBot. The punctuated lines indicate electrical
connections for communication.
The HCC is responsible for performing the HortiBot basic tasks such as position estimation, path
following control, payload handling, emergency response, etc. The HCC is an embedded
computer based on the industrial standard PC/104 architecture.
The vision module from Eco-Dan A/S, Denmark, is a new stereo vision system which captures
color and 3D information from horticultural and agricultural scenes. The output from the latter
system is expected to be adequate for the HortiBot navigating within transplanted onion parcels.
The standard transmitter or manual control unit for the Spider ILD01 slope mower is used as
remote control for the HortiBot. However, the Spider receiver unit, which also functions as the
Spider main control units, has been exchanged with a tailored CANbus enabled receiver,
Receiver-R-CAN NANO-L/A2, from NBB Germany.
8
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
3.2 Software
The main software solutions with concern to the HCC and the AVR based function modules will
briefly be presented in the following.
3.2.1 HortiBot Control Computer
The operating system of the HCC is an embedded Linux distribution, iComLinux developed by
Cetus, Denmark (www.cetus.dk). The iComLinux mounts the Compact Flash card read-only, and
during normal operations all writing operations are performed on a RAM-disk. This has the
advantage that the HCC can be switched off at any time without causing file system errors.
The HCC is connected to the sensors, actuators and communication interfaces via external
modules interfacing to the HCC via a Controller Area Network (CAN) bus or via serial ports.
The software architecture of the HCC is structured as a set of software modules interfacing to
each other via a shared data structure. Each software module is compiled as a Linux program,
and it uses the built in Linux shared memory and semaphore features to access the shared data
structure. Hence, the software modules can be started, stopped, added and upgraded
independently.
3.2.2 AVR Based Function Modules
In order to ensure a functional and stable design adaptable to future changes and functionalities,
the HortiBot design has been inspired by the automobile industry, which has a long experience in
creating stable and modulated designs.
This design provides the following benefits:
• Each function module handles all the detailed control of the individual functions.
• Each function module can be designed and tested independently of each other and the
HCC.
• Function modules can easily be reused in future applications. It is easy to make special
versions of modules to meet specific needs.
• It is possible to select the best computer/controller hardware in each module to obtain the
specific functions of the module.
• The total functionality of the HortiBot can be extended without being limited by the
capacity of the HCC.
• The benefit of a structured modularized design includes, that the demand for special
hardware for the HCC is dramatically reduced and a module can easily be changed
without any influence on other modules.
The Function modules have a shared design. All function modules require the following
components as a minimum (fig. 3):
• CAN-Protocol Handler handles the CAN-bus protocol.
9
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
• Command Handler is a component built to interpret the commands sent from the HCC via
the CAN-Protocol Handler and control the functions in the module.
• Module Function x (x = 1,…,n) are components handling the specific functionality of the
module and handles the interface to sensors and actuators used in the module.
• Utility Package is a package containing a number of general utility components. These
utilities will be available in all function modules, and be based on a joint source code.
CAN-Protocol Handler Command Handler
Module Function 1
Module Function n
Utility-
Package
Figure 3. The components in function modules.
3.3 The Voice of the Customer
Possible customer requirements were identified using various information sources like literature
review, current research activities in the robotic area, existing product screening, etc. Also, semi-
structured interviews with progressive horticulturists were used to consolidate the preliminary
requirement identifications. See also Sørensen et al. (2006).
Based on the modified importance ratings and the resulting importance ratings, the overall range
of requirements were sorted in descending order in figure 4. Important user requirements include
adjustability to row distance and parcel size, profitability, minimum damage to crops, and
reliability. Lower ratings are attributed to requirements like affection value, prestige; attractive
look, out of season operations, and use of renewable energy. The yellow full line in figure 4
represents the performance ratings of the HortiBot, the green line with punctuation represents the
importance ratings of the GreenTrac, and the black dotted line with punctuation represents the
tractor with AutoFarm AutoSteer. Score 0 equals to not important and score 5 equals to very
important. The result is based on 35 interviews made in Denmark, Germany, and Switzerland.
10
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
00.5 11.5 22.5 33.5 44.5
Effective
WorkAlone
ReduceManHours
EasyMount
EasyTransport
MinorService
EasyStartJob
NoShortStops
Reliable
EasyOperate
EasyService
Upgradeable
Flexible
CarryImplements
OperateByUnskilled
OperateSoftSoil
AutomaticDataAcquisition
OperateOutOfSeason
ReduceHeadland
NoHumanDamage
NoGrothMediumDamage
MinimizeCropDamage
Profitable
LowPurchasePrice
LowOperatingCosts
FastDepreciation
LowEnergyConsumption
ComparativelyQuit
ReduceRepetitiveWork
UseRenewableEnergy
AdjustableToRowDistance
LookAttractive
AffectionValue
LightWeight
SmallSize
Score
WorkCapacity
Function
Damage
Economy
Environment
Design
Figure 4. Average importance ratings for the overall range of requirements.
3.4 Competitive Tool Carrier Systems for Weed Control
In order to evaluate the market for Horticultural tool carriers in terms of identifying the relative
position of the proposed product (HortiBot) in the market and specifically, assign priorities for
further improvement, the already identified customers rated the relative performance of the three
competitive products using a 5 point score scale.
The overall performance ratings of the three competing products in figure 4 show that, for
example, the tractor with auto steering scores high on requirements like reliability, adjustability
to field conditions, effectiveness, flexibility, etc., while the GreenTrac scores high on
requirements like low energy consumption, automatic data acquisition, noiseless operation, use
of renewable energy, etc. The HortiBot gets high performance ratings on requirements like
reduced man-hours, minimized crop damage, profitability, reduced repetitive work, low
operating costs, easy to operate, etc. It is characteristic that, for example, the GreenTrac gets
relatively high performance ratings on requirements, which, on the other hand, the users deems
less important.
4. DISCUSSION
By modifying a remote controlled slope mower, it has been shown that it is possible to produce a
robust horticultural tool carrier. Weeding is the most profitable operation to automate within
outdoor horticulture. Nørremark et al (2006) concluded that most promising weeding tools were
HortiBot
GreenTrac
Tractor
(AutoFarm)
11
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
laser, rotary steel rods or L-tines and mower. All these tools are relatively light in their
construction. Hence, the HortiBot, which can only be able to carry relatively light implements,
seems a suitable carrier.
It was shown, that the most important user requirements attained to a robotic weeding tool carrier
include easy adaptation of the carrier to field conditions in terms or row distance and parcel size.
The HortiBot does not fulfill these demands entirely. However, the modular design makes it
relatively simple to adapt the HortiBot. Still, this will demand a redesign of the Spider ILD01
slope mower.
Due to the Open Source principles with concern to the HCC and the AVR based function
modules, the HortiBot may be of value for educational institutions and universities in need for a
simple and robust tool carrier.
5. CONCLUSIONS
By modifying a remote controlled slope mower it is possible to produce a robust horticultural
tool carrier for outdoor horticultural weeding. Due to the open source principles used, the
HortiBot may be developed further by other institutions.
QFD is a valuable tool that can be used when developing a new product. It is a structured method
where customer requirements can be analyzed and built in during the design stage. In this paper,
it was demonstrated how a selected part of the QFD process was carried out for a robotic tool
carrier to be used in horticulture.
Based on a combination of importance ratings and competitive priority ratings important user
requirements include easy adaptation to field conditions in terms of row distance and parcel size,
profitability, minimum crop damage during operation, and reliability. Lesser importance was
attributed to affection value, attractive look, the possibility of out of season usage, and the use of
renewable energy.
The study has demonstrated the feasibility of applying a systematic planning technique for
translation of the “voice of the customer” into the specific design and technical specifications of
a robotic tool carrier to be used in horticulture.
Further research will comprise identifying technical specification which best match the identified
customer requirements.
6. ACKNOWLEDGEMENTS
This research was supported by The Directorate for Food, Fisheries and Agri Business, Denmark
(DFFE). We are grateful to Director Lennart Ahlefeldt-Laurvig, Special Maskiner, Denmark; Dr.
Frank Hemmerich, KomTek, Germany; and Sales manager Jan Formánek, Dvořák Machine
Devision, Czech Republic, for technical assistance with concern to the Spider ILD01 slope
mower. The authors want to thank the participating horticultural users and also, the Swiss
Federal Research Station for Agricultural Economics and Engineering and the German
Association for Technology and Structures in Agriculture (KTBL), which have facilitated the
contact to users in these countries.
12
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
7. REFERENCES
Akao, Y. and G. H. Mazur. 2003. The leading edge in QFD: past, present and future.
International Journal of Quality & Reliability Management 20(1): 20-35.
Åstrand, B. and A. J. Baerveldt. 2002. An agricultural mobile robot with vision-based perception
for mechanical weed control. Autonomous Robots 13(1):21-35.
Callaghan, V., P. Chernett, M. Colley, T. Lawson, J. Standeven, M. C. West and M. Ragget.
1997. Automating agricultural vehicles. Industrial Robots 4(5): 364-369.
Chan, L. K., and M. L. Wu. 2005. A systematic approach to quality function deployment with a
full illustrative example. Omega 33: 119-139.
Cohen, L. 1995. Quality Function Deployment. How to Make QFD Work for You. Engineering
process improvement series. Addison-Wesley Publishing Company.
Crowe, T. J., and C. C. Cheng. 1996. Using quality function deployment in manufacturing
strategic planning. International Journal of Operations & Production Management 16(4):
35-48.
Edan, Y and A. Bechar. 1998. Multi-purpose agricultural robot. In: Proceeding of The Sixth
IASTED International Conference, Robotics And Manufacturing, 205-212. Banff, Canada.
Hague T., J. A. Marchant, and N. D. Tillett. 1997. Autonomous robot navigation for precision
horticulture. In: IEEE Int. Conf. On Robotics and Automation. 1880-1885. Alburquerque,
New Mexico.
Jørgensen, R. N. 2005. Præcisionsjordbrug – hvor svært kan det være?. In: Plantekongres 2005,
74-75, ISBN 87-984996-6-1. (In Danish).
Jørgensen, R.N., C.G. Sørensen, K. Jensen, H.T. Søgaard, J.M. Pedersen, L.B. Sørensen, I.
Havn, S. Sibbesen, and J. Nielsen. 2006. HortiBot: An Accessory Kit Transforming a Slope
Mower into a Robotic Tool Carrier for High-Tech Plant Nursing - Part I. In 2006 ASABE
Annual International Meeting, Portland Convention Center, Portland, Oregon, 9 - 12 July
2006.
Kassler, M. 2001. Agricultural Automation in the new Millennium. Computers and Electronics
in Agriculture. 30(1-3): 237-240.
Melander, B., and G. Rasmussen. 2001. Effects of cultural methods and physical weed control on
intrarow weed numbers, manual weeding and marketable yield in direct-sown leek and bulb
onion. Weed Research, 41: 491-508.
Nørremark, M., C.G. Sørensen, and R.N. Jørgensen. 2006. HortiBot: Comparison of present and
future phyto-technologies for weed control. In 2006 ASABE Annual International Meeting,
Portland Convention Center, Portland, Oregon, 9 - 12 July 2006.
Ørum, J.E., and J. Christensen. 2001. Produktionsøkonomiske analyser af mulighederne for en
reduceret pesticidanvendelse i dansk gartneri. Institute of Food and Resource Economics,
Denmark. Report no. 128.
Sørensen, C.G., R.N. Jørgensen, J.M. Pedersen, and M. Nørremark. 2006. HortiBot: Application
of Quality Function Deployment (QFD) Method for Horticultural Robotic Tool Carrier
Design Planning - Part II. In 2006 ASABE Annual International Meeting, Portland
Convention Center, Portland, Oregon, 9 - 12 July 2006.
13
R.N. Jørgensen, C.G. Sørensen, J. Maagaard, I. Havn, K. Jensen, H.T. Søgaard, and L.B.
Sørensen. “HortiBot: A System Design of a Robotic Tool Carrier for High-tech Plant Nursing”.
Agricultural Engineering International: the CIGR Ejournal. Manuscript ATOE 07 006. Vol. IX.
July, 2007.
Sørensen, M., and T. Frederiksen. 2002. Greentrac – Et miljørigtigt alternativ? Unpublished
B.Sc. (Hons) Vitus Bering, CVU, Denmark. Hovedrapport PRO M2. (In Danish).
