KITTING PERFORMED BY A COLLABORATIVE ROBOT WITH AN ANTHROPOMORPHIC TOOL
Master Degree Project in Virtual Product Realization One year Level 18 ECTS
Spring term 2020-07-02 Marcus Gottschlich
Supervisors: Niklas Land, PhD Student
Examiner: Anna Syberfeldt, Professor
Abstract
In the fields of automated manufacturing, flexibility has been a long struggle for companies. This factor is profoundly dependable on the demands of the customer in the form of high-quality products, variations, and low-prices, which has a high impact on the manufacturing industry. With the help of an anthropomorphic tool, the flexibility factor in, some extend, be the answer for companies working with robotic automation manufacturing
This thesis aims to investigate what kinds of objects that can be picked and placed, with the help of an anthropomorphic tool, mounted on a collaborative, UR3 robot—thereby evaluating the potential of what an anthropomorphic tool can and cannot pick. Also, investigate pick and place operation with different geometrical shape objects.
The project consists of identifying objects that are being picked and placed on a tailormade kitting box located on a conveyor belt. With help from the next station regarding what kinds of objects that are being delivered, a kitting box was designed and created. Stands could then be created for the kitting box and the objects. This was performed with the help of the software Creo Parametric, which was then 3D-printed. A shelf was created to successfully pick objects from, and thereby placing objects into the kitting box.
In the thesis, investigation regarding what kinds of objects the tool can pick was performed. This was conducted by picking the objects used for the thesis and everyday objects, such as duct tape, knife, and coffee cup. The experiment consisted of 15 different objects, were different grasping techniques were tested and how much force the tool used.
The result of the station as a whole has successfully been created in Assar innovation arena. The an- thropomorphic tool proved to be a concept to conquer the flexibility struggle in the form of success- fully grasping and placing different objects that have different geometrical shapes and weights. With the help of a conveyor, the whole station can also work in reverse mode, by picking objects from the kit-ting box and placing them into the stands, assembled on a shelf.
Certificate of Authenticity
Submitted by Marcus Gottschlich to the University of Skövde as a Master's Degree Thesis at the School of Engineering.
I certify that all material in this Master Thesis Project, which is not my work, has been appropriately referenced.
Signature.
Marcus Gottschlich
Table of Contents
1 Introduction ... 7
1.1 Problem description ... 7
1.2 Purpose ... 8
1.3 Objectives ... 8
1.4 Scope and Delimitation ... 9
2 Sustainability ... 10
2.1 Economic sustainability ... 10
2.2 Social sustainability ... 11
2.3 Ecological sustainability ... 11
2.4 Sustainability and its relevance to the thesis ... 11
3 The Frame of Reference ... 13
3.1 Standards ... 13
3.2 Robots ... 14
3.3 Cobots ... 14
3.4 QB Softhand research ... 15
3.5 Material Handling ... 18
3.6 Automated Production Lines ... 19
4 Literature study ... 20
4.1 Pick and place ... 20
4.2 Robot tools ... 21
4.3 Uncanny Valley ... 22
4.4 Sustainability regarding e-waste ... 23
4.5 Evaluation of literature study ... 24
5 Research methodology ... 26
5.1 Design & Creation ... 26
5.2 Observations ... 28
5.2.1 Systematic observation ... 28
5.2.2 Participant observation... 28
5.3 Documents ... 29
5.4 Choice of research strategy ... 29
6 Implementation ... 31
6.1 Research process & design planning ... 31
6.1.1 Suggestion step ... 32
6.1.2 Development ... 32
6.1.3 Evaluation ... 34
7 Results ... 36
7.1 Station ... 36
7.1.1 The shelf... 36
7.1.2 The pallet... 37
7.1.3 Flow-chart of the robot programming ... 39
7.2 Experiments ... 42
8 Discussion ... 44
9 Conclusion ... 46
9.1 Evaluation of the purpose and the goal... 46
9.2 Future work ... 48
10 References ... 49
Appendix A: Different widgets delivered to next station ... 51
Appendix B: Different robot structures ... 52
Appendix C: Coffee cup ... 53
Appendix D: A knife ... 54
Appendix E: A tennis ball... 55
Table of Figures
Figure 1, Author created visual integration between the three aspects ... 10
Figure 2, QB Softhand Research robotic tool (Qb SoftHand Research - anthropomorphic robotic hand – qbro- botics, 2020) ... 16
Figure 3, Model of the uncanny valley, inspired by (Rosenthal-von der Pütten and Krämer, 2014) ... 17
Figure 4, Model of the uncanny valley with moving and still objects, inspired by (Rosenthal-von der Pütten and Krämer, 2014) ... 18
Figure 5, Proportions with grippers, considering cost and flexibility, inspired by (Malik and Bilberg, 2019) . 22 Figure 6, Design science research process model inspired by (van der Merwe, Gerber, and Smuts, 2017) ... 27
Figure 7, Author own creation of process diagram of the thesis research process ... 32
Figure 8, Station layout ... 33
Figure 9, Widgets to be delivered ... 34
Figure 10, Pre-determined template ... 35
Figure 11, Final layout of the station ... 36
Figure 12, The shelf ... 37
Figure 13, Virtual model of the kitting box ... 38
Figure 14, The pallet ... 39
Figure 15, Flow-chart of the whole program ... 40
Figure 16, flow-chart considering reverse mode ... 41
Figure 17, Station ... 42
Figure 18, Experiments conducted ... 43
Figure 19, Grasping techniques QB Softhand user guide (2020) ... 43
1 Introduction
The introduction chapter consists of introducing the description of problems identified briefly. The chapter also includes, the purpose of the thesis, set objectives, scope, and delimitations.
1.1 Problem description
In the modern era, the manufacturing process has been improved and optimized. One contributing factor to the changes in manufacturing is the tools of automation, such as robotics, PLC, and simulation software. Mateus et al. (2017) state that, a struggle that still exists in the manufacturing industry is the unergonomic, heavy lifting tasks, and variations in the final assembly. These factors are common in the automobile industry, mainly whenever a truck engine is being produced and assembled. In some areas, some factors are hard to automate, some task that requires human abilities which includes fi- nesse, flexibility, and hard to reach jobs.
Operations regarding pick and place tasks is still an issue for most companies; whenever it comes to ergonomically, heavy lifting, or repetitive tasks, human labor is significantly impacted. According to Daria et al. (2017), warehouses that have to pick and place operations is estimated at 55% of the whole expenses of the warehouse. Even flexibility regarding pick and place operation is an issue due to the fact of human error. Pick and place operations do, in some cases, requires finesse and human touch.
However, collaborative robots have proved to be able to relieve human unergonomic and massive lifting operations.
An expedient that can provide for relieving such as heavy lifting tasks is the cobots, which stands for collaborative robots. Cobots can work reliably with humans to perform a specific task, such as heavy lifting or unergonomic elements. Different scientific papers and research pose favorable to collabora- tive robots.
A challenge for industry 4.0 is that it requires high flexibility in the manufacturing plant. According to Naseraldin, Chadhuri, and Pilati (2019), flexibility is a huge factor for the industry, for instance, it can be defined as several factors such as a high number of variants or product variations. So far, the human ability, regarding flexibility, can be seen as superior. It is challenging to choose a tool for tasks that requires flexibility and variations in geometrical product shapes. A tool is, in most cases, desig- nated for being used for a specific purpose and a specific product. If the company decides to produce a new product, the already bought tool might not apply to the new product.
The thesis will be investigating in, by using an anthropomorphic tool, called QB Softhand Research, manufactured by the company QB robotics. According to NE.se (2020), synonym anthropomorphic means that something is of human figuration. In this case, the QB Softhand Research has a figuration of a human hand.
With the help of the anthropomorphic tool, an investigation will be performed to examine if it is pos- sible to use the tool for pick and place of different geometrical widgets in a kitting operation. Regarding tools that have human figuration, there is little research on the area. The main reason for this could be that anthropomorphic tools are pretty newly released on the market, and therefore it is highly expensive robotic tools. The study that is being performed is of high relevance to new technology due to the small amount of existing research. Also, to investigate if the flexibility of the tool has a resemblance of a human hand can achieve the same flexibility as a worker's hand.
1.2 Purpose
The purpose of the thesis is to investigate if it is possible to use an anthropomorphic tool on a collab- orative robot that can perform tasks in the form of pick and place of different widgets that has different geometrical shapes. Thereby investigate the flexibility that may be achieved by using an anthropo- morphic tool.
This will also work with future purpose, that with the help of a kitting operation which is able to work reverse. To further prove that the purpose is achievable, a demonstrator will be built on Assar Industri- al Innovation Arena.
1.3 Objectives
Objectives are constructed in the main objective and sub-objectives. The main objective is the primary goal with the thesis, and with the help of the sub-objective, the main objective can be achieved.
The main objective consists of:
• Investigate the possibilities and flexibility with an anthropomorphic tool, with the help of a UR3 robot and pick and place operations.
The sub-objectives consist of:
• Introduce an anthropomorphic tool that shall be able to pick and place desired widgets to pre- define locations accurately.
• Using an anthropomorphic tool to prove that it is possible to pick and place different geomet- rical widgets.
• Investigate the level of flexibility regarding the chosen anthropomorphic tool by conducting experiments of different geometrical shape widgets.
• With inspiration from real industrial scenarios, create a tailormade kitting box with the desired widgets requested from the next station.
• The robot will be able to work reverse. Taking widgets from a kitting box and put them back to pre-define locations.
• Create a demonstrator to show the possibilities with the anthropomorphic tool for pick and place operations.
1.4 Scope and Delimitation
The focus of the project will be on kitting with an anthropomorphic robot tool. The task consists only of a robot that shall perform specific elements and is automated. The cell does not include a human- robot collaboration. There will be safety aspects that need to be implemented with the robot, such as safety plains to prevent the robot from reach outside its work area and following ISO regulations re- garding collaborative robots. Further limitations of the project consist of:
• The project is a station in a whole manufacturing line, consist of three other stations.
• It is noted that a part of the project is dependable on the transportation of the kitting box and the next station in the manufacturing line.
• Safety regulations regarding collaborative robots need to be investigated to create a safe envi- ronmental layout only for the UR3 station.
2 Sustainability
Sustainability can be defined as sustainable development, and according to Andrews & Granath (2016), it is a development that satisfies today's generations' needs without jeopardizing the ability of future generations to meet their needs. Sustainable development is classified into three different fac- tors, economic, social, and ecological sustainability. These three factors shall together work as a bal- ance with each other and thereby form sustainable development. An illustration of how the three as- pects are integrated into each other is illustrated in figure 1.
Figure 1, Author created visual integration between the three aspects
In general, sustainable development is using measurement as capital. According to Gullikson &
Holmgren (2018), this is categorized into four different factors, real capital, natural capital, social cap- ital, and human capital. Real capital is defined as what humans have produced, such as cars, buildings, and cloths. The natural capital is assets that nature provided, both renewable and non-renewable. Ex- amples of this can be oil, forest, and uranium. Social capital is defined as the culture and norms of civilization. Human capital is the ability of what humans in civilization can create and change in soci- ety.
2.1 Economic sustainability
The economic aspect is defined as not trying to achieve economic profit by overexploiting natural resources. Dahlin (2014) explains that this means, in simple terms, that the final value of a product
shall not over exciding the capacity of the natural resources. Economic sustainability does not neces- sarily mean that the currency is measured in money. It can be measured in the form of raw material, energy, or work hours. In order to achieve economic sustainability, long-term thinking is required, such as making an economic profit in the long term instead of short term.
2.2 Social sustainability
The definition of social development is focused on society and human rights. According to Dahlin (2014), the aim is to achieve a long-term society, meaning that social sustainability is aiming for human rights, safety, freedom, reduction of poverty, among many aspects. Social sustainability also includes humans as a whole, which means both physical and mental. It is essential that humans have a right to privacy and can be heard and seen in their working life.
2.3 Ecological sustainability
For nature to recreate the natural resources that have been overexploited, an ecological approach is necessary to prevent this kind of behavior. Dahlin (2014) states that, in order for nature to recreate the natural resources that have been overexploited, an ecological approach is necessary to prevent this overexploiting. The cycle of nature is only able to absorb a certain amount of pollution, and it takes a long time to recreate natural resources. Thereby it is of high importance to limit the pollution from an industrial manufacturing plant and not overexploiting resources.
2.4 Sustainability and its relevance to the thesis
In this thesis, sustainability has a material impact on all three factors. Regarding the economic aspect, the anthropomorphic tool that is going to be used is working almost like a human hand. One factor with this anthropomorphic tool is that it is still an expensive robotic tool. This kind of tool is still on a research-level and is not implemented in many factories.
Regarding the impact on the ecological factor, as mentioned, this anthropomorphic tool is still on a re- search-level, and the production of these tools is limited. It can be produced whenever a request from the customer is being made. The manufacturer of the tool is located in Italy and has to follow EU- directives, thereby ensuring ecological aspects such as unnecessary pollution. With the use of this tool and automation, it might be proved that the tool can provide for lesser quality errors. It is thereby reducing the ecological impact by not discarding the product.
The last aspect is the social factor. The tool can be used for additional payload for laborers in the form of unergonomic movements, also facilitate repetitive tasks for the operator. By using a tool that has a look-alike appearance of a human hand, flexibility can be achieved due to the fact that labors can be seen as the most flexible resource. In this way, the tool can perform a similar task or sharing a task with a human. Also, the tool can provide for eliminating sick leave due to injuries, working in a man- ufacturing line.
3 The Frame of Reference
In this chapter, the investigation will be executed regarding the relevant elements of the project. The investigation consists primarily of traditional robots, tools, collaborative robots, safety, and kitting elements. Whenever working with robots or general machinery, it is good to achieve an overview regarding safety standards and what it is in order to work safely. The reason for investigating robots, cobots, and UR robots is to achieve general knowledge and perspective of them. This leads to the tool of the robot, primarily the chosen anthropomorphic tool, the QB Softhand, were overall knowledge about the tool needs to be achieved in order to work correctly with it. Overall the project consists of using the anthropomorphic tool to provide products to the next station, which is the reason material handling is be-ing investigated in order to achieve knowledge of how to work with that kind of process.
The station shall be fully automated, and this means that automated production lines will be investi- gated in, thereby achieving knowledge of how this kind of system works and what to think about when working with it.
3.1 Standards
According to the Svenska Institutet för Standarder, SIS (2020), a standard is a mutual agreement re- garding solutions to recurring problems. Standards exist in all areas such as small nails, screws, data communication, industry, healthcare, and environment. One can make every part of the existence af- fected by standards. At the Swedish International Standards, there are national and international stand- ards that have all been anchored developed and quality assured through the standardization process.
Standards have been developed in a variety of industries and areas. It is voluntary to apply for stand- ards; however, it can serve as a mandatory reference, for instance, regulations from pictures and pro- curement.
CE marking
According to SIS (2020), CE marking is the abbreviation for Conformité Européenne, which means following EC directives. Whenever there is a CE mark on a product, then the conclusion is that the manufacturer or the importer certifies that it complies with the European Union health, environmental, and safety requirements. It is also a trademark, which means that a CE marked product can be sold freely within the European Union. To know what applies to each product, one needs the regulations that transfer the directives. The requirements that the European Union directive has, also transfer to Swedish rules. The operating instructions that come with the product that is, for instance, transferred
to Sweden are, therefore, in Swedish. The corresponding rules are applied in every country within the European Union.
3.2 Robots
An industrial robot is according to Industrial Robots (2019) defined through the ISO 8372:2012 "An automatically controlled, reprogrammable, multipurpose manipulator programmable in three or more axes, which can be either fixed in place or mobile for use in industrial automation applications." For clarification from ISO 8372:2012, a reprogrammable option means that it will be possible to reprogram the motions without making a physical change. A multipurpose option means that the robot will be able to change an application by changing the physical structure.
According to Industrial Robots (2019), by the year 2004, an agreement was set regarding industrial robots that consisted of a robot will be defined as a mechanical structure. There are a lot of different kinds of robots regarding change of mechanical structure; see appendix A. For instance, a cartesian robot consists of three different prismatic joints, where the axis is in correlation with a cartesian coor- dinated system. Another example is the articulated robot that has an arm, consist of least three different joints; see appendix A for illustration.
3.3 Cobots
Cobots or collaborative robot is a robot that can work together with a human in a safe way. Whenever a company or a factory wants to implement a collaborative robot, risk analysis needs to be performed, thereby ensure the human aspect. The risk analysis consists of evaluating, was the robot integrating at, such as areas and tasks. During this stage, these tasks can be evaluated and thereby minimize or elim- inate injuries for the worker. Some safety aspects can be complemented with the help of ISO standards.
(SIS-ISO/TS 15066:2016) UR-Robots
UR-robots or Universal robots is a specific company that manufactures collaborative robots. Accord- ing to UR (2020) The universal robots consist of six axes and are designated to collaborate with human workers. Universal robots have been developed two types of UR-robots, the traditional CB-series, which came first and the latest e-series. The e-series have some additional features, such as inbuilt power and element sensors. The company has four different robots that is out on the market, which are UR3e, UR5e, UR10e, and UR16e. Differences between the four models depends on reachability and
payload, were the UR3e has the least payload, and the UR16e has the highest. This also means that the UR3 has a payload of a maximum of 3 kilograms, and the UR16 has a payload of a maximum of 16 kilograms. As seen, the last number in the respective robot stands for the maximum payload without equipped tools.
Stated by UR (2020), whenever the robots are implemented in a factory, they can work with tasks such as pick and place, assembly, packetizing, quality checking, and screw-drawing. There are five different advantages by using UR-robots; the installation of a UR-robot is fast and easy; it is easy to program through online-programming, UR-robots have high flexibility, and finally, the UR-robots tend to repay themselves fast. The UR-robots have several safety functions and can be equipped with several differ- ent grippers and tools, which can also be tailormade for a specific purpose.
3.4 QB Softhand research
When it comes to the choice of an anthropomorphic tool, the QB Softhand research fits the configura- tion for the designated task. Qb SoftHand Research - anthropomorphic robotic hand - qbrobotics (2020) clearly states the soft hand is a flexible, adaptable, and that has the ability to interact with the environment. It is mentioned that the tool has a minimized risk for hurting humans or the robot itself.
The tool has been configurated were flexibility perspective is much considered. It is stated, according to Qb SoftHand Research - anthropomorphic robotic hand - qbrobotics (2020), that the tool does not require any changes in the control panel when grabbing different objects. The tool can be visualized in figure 2.
Figure 2, QB Softhand Research robotic tool (Qb SoftHand Research - anthropomorphic robotic hand – qbro-botics, 2020)
This tool is only meant for research and educational purpose and therefore not suitable for the factory.
According to QB Softhand Research User Guide (2019), the tool offers a grasping range of objects within the range of one to 1700 grams. The tool shall not be used whenever working with objects that is fragile or has sharp edges. The user guide also states that objects that have lesser or equal to eight- millimeter shall not be grasped, also objects more significant than 300 millimeters shall not be grasped.
The tool offers a 19 degree of freedom, has a grasping force of 62 Newton, and the payload is maxim- ized to 1,7 kilograms.
Uncanny valley
The concept, uncanny valley was, according to Rosenthal-von der Pütten and Krämer (2014), a the- ory that was developed in 1970. It is stated that the uncanny valley can be defined as when people are interacting, a correlation of a linear function, and the relationship between two random variables is described. Whenever people are interacting with something that cannot be explained by the linear function or one of these variables' shatters, confusion emerges.
In the fields of robotics, it is stated by Rosenthal-von der Pütten and Krämer (2014) that whenever individuals are trying to make robots appear more human-like, it also leads to an increase in the af- finity level, regarding the development of human alike robots. The affinity level is defined as hu- mans feel more vicinity or closeness to the object in general. The function leads to increment until it reaches the uncanny valley, where individuals feel discussed and uncomfortable with the object.
Figure 3 demonstrates the uncanny valley and how the affinity level is being affected. For example, it demonstrates how industrial robots affect the affinity level, where it is a little higher than zero. On the human likeness axis, the industrial robots are also not high, thereby the repulsiveness is not com- mon for industrial robots. The prosthetic hand is close to 100 percent and the affinity level is low, which can be described carefully to repulsiveness from a human perspective.
Figure 3, Model of the uncanny valley, inspired by (Rosenthal-von der Pütten and Krämer, 2014)
Figure 4 illustrates the correlation between moving objects and still objects. The filled curve can be a representative mirror of the curve. The dotted curve is whenever the objects are moving, such as an operational humanoid robot. It is visualized that at the bottom of the dotted curve, zombies are one of the factors that, according to the uncanny curve is most extremely repulsive. It is also noted that a myoelectrical hand can be defined as tightly to repulsive or uncomfortable, as it has a close appear- ance to an anthropomorphic hand.
Figure 4, Model of the uncanny valley with moving and still objects, inspired by (Rosenthal-von der Pütten and Krämer, 2014)
3.5 Material Handling
Material handling is, according to Freivalds & Niebel (2013), several factors that consist of time, place, motion, quantity, and space constraints. The factor motion is about having a regular flow of widgets, raw-material, final products, supplies, et cetera to different locations in a factory.
Time is an essential factor that consists of material handling. According to Freivalds & Niebel (2013), the time factor states that each operation is requesting material at a particular time and thereby ensures that the process is not being late, or that customer is suffering from delayed products. The place factor refers to deliver the correct amount of material to the station or operation that requests the material.
According to Freivalds & Niebel (2013), the factor quantity state that the materials that are being delivered to the designated station shall not be damaged or defect. The last factor, space constraints, states that the material must have both temporary and passively room for storage space.
Kitting
According to Groover (2015), a kitting operation is whenever parts or widgets are being stored for operations such as assembly or disassembly. When an order is being received for the kitting station, the demanded products or widgets are collected or retrieved into a kit or a tote pan. After the station has collected the required widgets into a kit, the station shall deliver the kit to the station that requested the kit.
3.6 Automated Production Lines
Whenever a design of automated production lines is about to be performed, then there are, according to Groover (2015), three different aspects that need to be considered. The first aspect is line balancing.
This factor can be considered as a central problem in the manual assembly line; however, it is also a problem in the automated production line. Line balancing is in proportion to optimization of automated production line. To balance the line, consider working with automation, the line is constraint by tech- nology advancement and needs to be fulfilled in specific steps. For instance, line balancing is also about creating a balanced work-flow, were the first station does not need to wait for the next station to finish its process and vice versa. The second consideration is the processing technology, and it states the knowledge required or achieved for a specific manufacturing process to work.
It is mentioned that, according to Groover (2015), for a specific manufacturing process to be imple- mented, a lot of investigations needs to be conducted. For instance, a lathe machine needs information regarding cutting tools, speed, chip control, selection of material, et cetera. Implement manufacturing technology has been refined during many years of practice and research; thereby, one can choose to set the settings of a machine to its maximum capacity performance. Third and lastly, consideration is the system reliability; one needs to consider when working with automated lines, a complex system can occur. This means that when a failure of a component occurs, such as, a conveyor belt the whole production line can stop. With the help of a mathematical model to calculate, for instance, cycle time, this problem can be prevented.
4 Literature study
This chapter describes some relevant literature that has been conducted and performed in real life. The literature that is presented in this chapter has high relevance to the project that is being performed. The articles that are being investigated on consists of pick and place, collaborative robots, robotic tools, convolutional neural networks, flexibility, the uncanny valley, and sustainability.
4.1 Pick and place
Learning robots to grasp by demonstration (De Coninck et al., 2020)
The paper Learning robots to grasp by demonstration, written by De Coninck et al. (2020), discuss programming of collaborative robots typically. In the paper, this approach of programming is called
"program from a demonstration." The authors are using a pick and place approach, and the concept builds on the manufacturing approach called "learning from demonstration." This means, according to the paper, that an operator manipulates the cobots to specific points or poses in space and thereby can replay those poses in a program.
The authors are using this kind of approach, however with usage of machine learning and given the feature from cobots, record and replay, the paper called its own study "program from a demonstration.
It is pointed in the paper that the authors notice a negative factor by using "program from demonstra- tion," which consists of cannot adapt to changes in the workspace.
The case is being performed by trying to simulate how big the grasp success rate is by using a convo- lutional neural network. In order to demonstrate how accurate, the convolutional neural network is, the paper is using of ordinary objects that consist of the duct tape roll, coffee mug, screwdriver, stapler, dispenser, wire clipper, and a whiteboard marker. The authors achieved a high success rate on most of the objects. Whenever the authors applied the collected datasheets on a real collaborative robot, a success rate of 91,34 percent was achieved with five seconds average to pick the objects.
Design and performance of kitting and order picking systems (Brynzér and Johansson, 1995)
The paper is written by Brynzér and Johansson (1995), discusses several different case studies regard- ing the performance and design of kitting systems. Specific, the paper investigates in picking order,
activity, work organization, method of picking, information system, and equipment. These factors are, according to the paper, in relation to measurement performance, such as efficiency and accuracy of picking.
According to Brynzér and Johansson (1995), a kitting operation means that a station provides other stations with components with kits. The reason for implementing a kitting operation is to ensure qual- ity, products that require high amounts of widgets, components of high value, or already implemented parallel assembly system. Whenever a redesign or design a new kitting system is about to be per- formed, some considerations regarding the design need to be considered. According to Brynzér and Johansson (1995), several factors need to be considered, such as orders in a batch, sequence widgets that need to be retrieved, and storage of widgets. The results, in the paper, is that improvements re- garding kitting systems can be efficient if a redesign of widgets or products can be achieved, consid- ering a kitting system that is in focus.
4.2 Robot tools
A robotic gripper for picking up two objects simultaneously (Atakuru and Samur, 2020)
The paper presented and written by Atakuru and Samur (2020) investigates the usage of a two degree of freedom gripper for pick and place tasks. This is performed due to overcome difficulties such as accuracy, precision, rigidity, and speed. The development of the case studies begins with design a gripper for the designated purpose. The purpose is to reduce cycle times and reduce energy consump- tion for the end effector, also referred to as a manipulator in the paper.
The authors Atakuru and Samur (2020) states that in the second phase consist of analyzing how a three-axis delta-type tool behave during a pick and place operation. The third phase of the case study consists of implement the gripper on the end effector and run with the same configuration as the delta gripper. The result, given by the paper, states that the design of the three-axis gripper performs its designated task with no problem, and it also reduces the cycle time compared to the traditional delta gripper.
Complexity-based task allocation in human-robot collaborative assembly (Malik and Bilberg, 2019)
Complexity-based task allocation in human-robot collaborative assembly written by the authors' Malik and Bilberg (2019) are discussing the increasing development in factories regarding skills by workers.
The authors are trying to present a methodology for distributing tasks, regarding assembly station be- tween a robot and human worker were complex tasks are being performed. The approach that the paper is presenting consist of using a collaborative robot for real cases and comparing to real industries cases.
The authors are presenting exciting facts about flexibility regarding the choice of tool. It is noted in the paper that there is a correlation between the choice of tool regarding flexibility and the final price of the tool, see figure 2. During the study and the approach of the methodology set by the authors, the conclusion is being set that the methodology can, in a way, differentiate tasks with higher complexity, safety, handling, et cetera.
Figure 5, Proportions with grippers, considering cost and flexibility, inspired by (Malik and Bilberg, 2019)
4.3 Uncanny Valley
How the design characteristics of robots determine evaluation and uncanny valley related re- sponses (Rosenthal-von der Pütten and Krämer, 2014)
The background of this article, written by Rosenthal-von der Pütten and Krämer (2014), consists of an investigation regarding the uncanny valley. The article presents an investigation where the author con- ducted a web-based survey and evaluated 40 different images of robots. These images consisted of robots that are not recognized on the marked by the general public, also the robots in the pictures varied in size, color, design, et cetera.
The participants that were evaluated when observing the pictures were, according to Rosenthal-von der Pütten and Krämer (2014), in accordance with the uncanny valley. The images were viewed for the participants with a questionnaire for the corresponding image. Whenever the images were viewed for the participants, facial expression was recorded.
The result from the paper indicates by the author Rosenthal-von der Pütten and Krämer (2014) that there are different categorize regarding the design of robots that are intimidating of humans. For in- stance, the result shows that height, color, facial expression, et cetera was profoundly dependable re- garding the evaluation test. An interesting aspect was that according to the participants, slender robots were more familiar, and the bulkier robots were more intimidating. A particular aspect that was affect- ing the result was bipedal design, which is a more terrestrial design, for instance, an ostrich.
4.4 Sustainability regarding e-waste
Collaborative robots in e-waste management (Alvarez-de-los-Mozos and Renteria, 2017)
Alvarez-de-los-Mozos and Renteria (2017) present and discuss in the paper that some industries have problems dealing with electronic waste or, as the article states, e-waste. The purpose of the article is to optimize the recycling of electronic devices or equipment, trying to implement criteria regarding economic and technical aspects. The paper shall also try to uncover waste considering the development of collaborative robots.
To fulfill the requirement, the authors try to implement a circular economy, thereby trying to apply a reverse logistic system. The system builds on the same concept as disassemble a product. In the article, statements are being set that traditional industrial robots require huge fences, time-consuming pro- gramming, and experienced programmers, thereby implementing a collaborative robot instead. One reason is, as the article states, collaborative robots allow close cooperation with human labor, thereby identification of material and components could rely on humans.
4.5 Evaluation of literature study
The articles that have been summarized in the chapter have relevance for the thesis as a whole. The first article consisted of learning cobots diffe4rent grasping techniques. This has high relevance to this project due to the programming of the chosen UR3 collaborative robot. Different grasping techniques were evaluated in the paper, such as “learning from demonstration” and “program from a demonstra- tion.” One of those techniques will be applied in the station during the programming of the collabora- tive robot.
The second article consisted of different case studies regarding the performance and design of a kitting box. This has high relevance for the project, due to the design and creation of a kitting box for widgets.
It is stated in the article that it is of high relevance to consider the batch size of the product and sequence of widgets, also the design of widgets. This will be into consideration whenever the creation of the kitting box takes place.
The third article that is being investigated is about an evaluation regarding three-axis gripper with two degrees of freedom. The reason for investigating this type of article is because of the nature of the chosen tool in this project. The anthropomorphic tool that is being used in this project is a five-finger tool with 19 degrees of freedom. In this case, the anthropomorphic tool has more gripping fingers and a lot of more degrees of freedom. However, the author in the article wants to investigate and overcome fac-tors, such as precision and accuracy, which is one of the sub-objectives with this thesis, by accurate pick and place widgets that have different geometrical shapes.
The fourth article is an investigation regarding an evaluation of different robotic grippers, which con- sist of a two-finger gripper, a three-finger gripper, and a five-finger gripper. The last gripper is of most interest for the purpose of this thesis. One of the sub-objectives with this project is to investigate the flexibility regarding the anthropomorphic tool. In this article, the authors present a table were the five- finger tool has the highest flexibility, which is in the correlation of this thesis set objectives.
The fifth article investigates the theory of the uncanny valley. This is being investigated due to the nature of the chosen tool. By using an anthropomorphic tool, individuals can achieve a disgusting feel- ing about the tool. The article states that a survey of different robots as several different appearances of humans. In the future, a survey could be applicable to this anthropomorphic tool in order to inves- tigate what individuals on a local level think about these kinds of tools.
The sixth article consisted of investigation regarding e-waste were the authors states that achieving a circular economy, or a reverse logic system more sustainable manufacturing can be achieved. This is of the high relevance of this thesis due to the objective of the station to be able to work reverse. It is also highly likely that one of the delivered widgets is electronic, thereby the article has high relevance of the project.
5 Research methodology
This chapter introduces and presents the chosen research methodology. Description of the research methodology is first described in general and how it is affecting the project in general. The chosen methodology regarding data collection shall also be described in this chapter.
5.1 Design & Creation
The investigated strategy is Design and creation. According to the book Researching information sys- tem and computing written by Oates (2006), design and creation is a research methodology that con- sists and using creative thinking and technical development work. This methodology differs from other strategies, such as instead of using theories that can have an abstract approach. Design and creation use the approach to end up creating, for instance, a physical product.
According to Oats (2006), the design and creation consist of following a five steps approach. This is an iterative process, and it uses a problem-solving approach. The design and creation methodology is illustrated in figure 7. The five steps consist of, and are presented in chronological order;
• Awareness: The first steps involve the user, or the researcher recognizes, identifying, or artic- ulate the problem itself. This can come from, for instance, identifying literature reviews, areas that require further research, or from newly discovered technical developments.
• Suggestion: The second step consists of a form of curiosity about the problem, thereby a leap from the state of curiosity to problem-solving —an offering of a preliminary perception of how the problem itself can be solved.
• Development: The third step consists of design generations and their implementations in order to reach a satisfactory solution. This approach is dependable on what kind of problem is ad- dressed. The solution is established in proportion to the addressed problem.
• Evaluation: The fourth step consists of an investigation regarding the generated design. The parameters that are being investigated regarding the generated design consist of, asses its val- ues, and if the expectations deviate.
• Conclusion: The fifth and last step consist of having a discussion with, for instance, other researchers, were the results are being noted and written. The knowledge that has been acquired during the process is identified together with, for instance, unexpected results that cannot be
explained or proven. Therefore, even more, research is perhaps necessary for further studying the anomaly.
Figure 6, Design science research process model inspired by (van der Merwe, Gerber, and Smuts, 2017)
By using and implement the methodology, design and creation, comes with a certain amount of ad- vantages. According to Oates (2006), the first advantage consists of a physical product or object that is able to be shown for the public or other researcher. Other scientists or researchers shall see and feel how much effort or work the user has contributed to the research. The second advantage is considered to individuals who enjoy or have a close interest in technical development. Thereby the technical il- lustration shall appeal to people who have a general technical interest. The third advantage is that the method as a whole is the expected approach for conducting research in some specific areas. This could be, for instance, the area of IT there the specialization may consist of software engineering or general computer science. The fourth advantage and last advantage, according to Oates (2006), is more of an advantage in the IT sector. Computers, in general, has not yet been implemented in some fields and continually increasing the performance of computers, much technical knowledge can be conducted or achieved in those domains.
5.2 Observations
According to Oats (2006), the fundamental definition of observation is first of all, a data collection technique, were the user uses, a watch and pay attention approach. Every individual uses this kind of approach in daily life, for instance, seeing, hearing, noting, forming theories, et cetera. Whenever a researcher uses the data collection technique of observation, it is often about the generation of data;
this could be, for instance, observe how individuals behave in a specific environment. This method can be applied in every research field, for instance, in the academic for research studies or the behav- iour of individuals during a meeting. Observation methodology can be a very structurally way of ap- proach, whenever conducting research. There is a clear case when observation deviates, and it is when- ever a researcher is conducting overt or covert research. In covert research, the individuals that are being observed do not know that the method is being performed. Overt research is the opposite of covert research, meaning that the affected individuals are notified about and when a researcher is con- ducting the technique. This data collection technique can have a time interval of five minutes to several years. The researcher or the user performing the observation data collection technique does not have an obligation to give feedback to the affected company or organization. The user chooses for them- selves to release no feedback or full feedback.
5.2.1 Systematic observation
According to Oates (2006), the definition regarding systematic observation is events that are prede- fined and is being investigated; the event is subjected to the observation. The systematic observation technique is making use of frequency and duration, and these factors are being used as counting and timing. To notify, these two parameters, a pre-designed schedule can be necessary to use. A researcher or a user who works with this method is following a systematic observation.
5.2.2 Participant observation
The participant observation is, according to Oates (2006), a form of observation where the user that is being affected by the method knows about the conduction. It can also be that the users are not informed about the method being performed. The researcher or the user performing the observation informs the affected individuals that observation is being performed. It can be possible for the researcher to still conduct an observation without informing the participant. The researcher can, therefore, trick the par- ticipant into informing the affected individuals that the researcher is a participant. The participant ob-
servation can be performed with the approach of either overt or covert observations. Even if the re- searcher or user chooses the overt or covert observation, it will be necessary to take as many notes as possible during the participant observation.
5.3 Documents
The method for generating data called, documents is, according to Oates (2006) consist of a source of data, another sort of approach regarding interviews, questionnaires, and observations. The method documents can be categorized into two different types; found documents and researcher-generated documents. The found documents are often used by the researcher, which consists of documents that are accessed in organizations or companies. In companies' documents can be represented as production schedules, job descriptions, profit, et cetera. The researcher-generated method is only meant for tasks regarding researching areas. An example is a researcher that conducts research involving taking pic- tures, thereby transform pictures into data by analyzing and observing. The pictures are becoming relevant for a process regarding illustration and justification in a design methodology. Documents method can be applied in every research strategy. A number of strategies were the method is applicable can be, for instance, design and creation, case studies, surveys, experiments, action research, et cetera.
An example regarding the design and creation strategy is that reports provided by a company are able to visualize what an end-user is expecting from, for instance, computer software or a computer system.
One significant source to make use of whenever working with a documental method is the internet.
Thereby making use of the internet when conducting research, electronic documentation is being per- formed. An example of when conducting the e-document method is whenever one is checking emails, performing personnel, and company webpages, online auctions, et cetera.
5.4 Choice of research strategy
The project aims to create a demonstrator with physical widgets; therefore, the design and creation methodology is the best-suited approach and the primary research strategy that will be followed. This approach is very suitable for engineers in general. The project, in general, can be challenging, pro- gramming a anthropomorphic robotic tool and create kitting boxes. According to Oates (2006), the design and creation cover many of parameters in order to conduct a project with this magnitude.
Observation and documentation are the data collection and data analysing techniques, that is being used in this project. The first one, observation, will be used; for instance, a meeting is being held. Since this project involves other projects, meetings will be held, and the importance of what is being decided
on these meetings shall take into consideration for this project. Documentation will be applied when- ever, for instance, the anthropomorphic robotic tool arrives and reading its manual. It will be of high importance to follow the companies guidelines whenever the installation of this tool takes place.
6 Implementation
In this chapter, suggestion regarding the implementation of the solution to the problem itself is present- ed. The research process and design planning and its fundamental parts are presented of how identified problems are being tackled together with developments of widgets and its stands.
6.1 Research process & design planning
In figure 8, the whole research process includes and centers around the research strategy design and creation. In each of the five steps, different aspects are being considered and will be conducted in relevance to the project itself. The starting point of the diagram is flexibility requirement, anthropo- morphic tool, and literature review. The flexibility requirement is an essential requirement due to the struggling industry faces every day with this factor. The literature review also confirms that flexibility in the industry is one of the biggest challenges. The anthropomorphic hand is a tool that can be a helpful robotic instrument to solve this problem.
In the first step, awareness and identification of the problem have been performed. In this case, the problem is that an anthropomorphic tool has hardly been used in the industry, and there is very little research on it. In the suggestion step, a hypothesis has been set on how to approach the problem and solve it. Due to the nature of the tool, the hypothesis consists of using it as a human hand, for instance, programming the tool to perform in the same way as a human hand would pick up a widget. In the third step, 3D-printed stands were created with the help of the software Creo. These stands will be tailormade, suitable for the required widgets. Also, the development step includes an experimental phase with the robotic tool in order to observe how the tool performs with different conditions. In the evaluation step, the testing of the program will be evaluated in order to observe if the tool or the robot collide in different objects and observe how well the robot pick up designated widgets. Lastly, the conclusion phase consists of creating a fundamental baseline for future work. This is performed in order to achieve another perspective of the ongoing program so that improvements can be performed.
This is also an iterative process, where improvements and changes will be performed in order to achieve a rigid and time-efficient workflow.
Figure 7, Author own creation of process diagram of the thesis research process
6.1.1 Suggestion step
In order to solve the identified problem, a robust suggestion needs to be achieved. An investigation regarding different already provided tools and equipment needs to be identified, in order to evaluate what kinds of tools that are available for usage. Tools and equipment that are used and identified in the station are a UR3 robot that includes a teach pendant, a newly bought anthropomorphic tool, work- table for the UR3 that has a dimension of 1000*700 mm and a conveyor belt.
In order to achieve knowledge regarding what kinds of objects the anthropomorphic tool shall grab, a discussion with the user in the next station was conducted. The identification was performed regarding measurement, geometrical shapes, and weight of the objects. These factors were included whenever the stands for the objects was virtual created in the CAD software, Creo Parametric.
6.1.2 Development
The whole project consists of creating a demonstrator on the Assar innovation arena. During this phase of the project, a number of inputs, both from the user of the next station and extracted data collection, were performed.
Demonstrator
The project consists of creating a demonstrator, located at Assar innovation arena. With the help of the tools in general and stands, creation could be performed. In order to achieve a robust layout of the station, a virtual layout was performed, see figure 9. The layout is a rough concept of how the robot, worktable, and the pallet will be placed.
Figure 8, Station layout
The kitting box was assembled in the software Creo Parametric. The pallet that is being transported on the conveyor acts as a vehicle. The measurement of the pallet was extracted and implemented in Creo, thereby having a virtual model of the pallet and therefore stands could be assembled on it. Several different virtual kitting boxes were created and illustrated for the user in the next station in order to achieve a mutual benefit and agreement of the virtual kitting box. For instance, a significant demand and constraint, requested from the next station were that the widget that is being picked by the robot should be on a specific side of the pallet. Several constraints have been identified, such as demands from the next station, the maximum length of the robot, a maximum payload of the robot, and the total area of the pallet. The pallet also has a designated spot on the conveyor, with the help of signals from a sensor, thereby achieving the same location of the kitting box whenever it arrives at the station.
The stands that were developed, together with insight from the user of the next station and Creo para- metric. The stands were tailormade and created to transport an electrical cabinet that consisted of a top and bottom plate, four M6 screws, and four different sensor outlets. This is illustrated in figure 9. These
widgets were performed with inspiration from a real Volvo truck engine. Regarding the widgets that are being delivered to the next station, a paper-thin business card was also delivered. This was per- formed in order to test the limits of the tool and to prove that the tool has a wide range of greppable objects, not only industrial equipment. The different widgets that are being delivered to the next station are illustrated in Appendix A together with indexes.
Figure 9, Widgets to be delivered
In order to pick up the widgets and place them onto the kitting box, a shelf was created. The shelf act like an inventory was all the different widgets is placed with the help of 3D printed stands. Whenever the pallet comes to the designated spot on the conveyor, the robot will start picking widgets from the shelf and place them onto the kitting box. The shelf is also mounted behind the robot on its worktable, and thereby the robot will work in an area that corresponds to more than 180 degrees workspace for the robot. The shelf needs to be adjusted further in the project in order for the robot to reach, both the shelf and the pallet.
6.1.3 Evaluation
Whenever the anthropomorphic tool was tested to pick different objects, observation during the pro- cess was performed. A pre-determined template was created with the aim of giving all the transparent in-formation, see figure 10. The sheet consisted of collect data in the form of grasping pressure in percent-age, tool poses, angle of the tool, et cetera. The reason for creating this kind of template is to make a transparency of the work in order to provide information to the general regarding what kind of objects the tool can grasp, which pose to use, grasping techniques, et cetera. Different objects have already been chosen for the experiments, which consist of mostly ordinary and daily objects, such as a bottle, duct tape, brush, and so et cetera. The reason for choosing ordinary objects is to investigates if the anthropo-morphic tool can grasp these kinds of objects with human functionality, considering
the tool only using one servomotor and 19 degrees of freedom. Not only using abstract industrial widg- ets such as sensor outlet or other electrical parts. Also, the idea is to give transparency of the work to individuals who not only have a technical background, thereby using these kinds of objects, a larger target group of individuals can relate to the tool.
Figure 10, Pre-determined template
7 Results
In this chapter, the results are being presented. In this section, results regarding the station as a whole and what has been conducted within the cell, such as the shelf, the pallet, and the robot programming.
The result chapter also includes experiments regarding objects et cetera.
7.1 Station
The final layout of the station is illustrated in figure 11, where the station can be seen as a whole from a birds-eye-view. The robot has been configured, in the form of location, to reach the shelf, located on the worktable, and the kitting box, located on the conveyor belt.
Figure 11, Final layout of the station
7.1.1 The shelf
The shelf was first located at the end of the robot worktable. However, this was changed so that the shelf was re-located closer to the robot for reachability purposes. The shelf has been re-located ap- proximately 200 mm. Otherwise, the robot would not be able to reach both the shelf and the pallet.
This is visualized in figure 12. The shelf is tailormade to suit designated purposes, which is picking different objects. It is made of wood, which has been painted grey for esthetical purposes. The wooden
plank has a dimension of 180*800 mm and is assembled on two different aluminum profiles, which has a dimension of 220*45*45 mm. This means that the shelf is standing on a height of 220 mm and is fixed onto the worktable of the robot. There are six different stands for the shelf, and these were 3D printed to achieve a fixed and stable position for the different objects. Drilling the shelf was performed to achieve a fixed position for the stands, and thereby screwing in the stands on the shelf was per- formed. The seventh stand is located and fixed on the worktable of the robot.
Figure 12, The shelf
7.1.2 The pallet
In order for flexible, prosperous, and stable transport of the different objects to the next station, a virtual kitting box was created in CREO Parametric, which is visualized in figure 13. The final model
was created with inputs from the user of the next station. The stands were manipulated in order to achieve a model that satisfies both the user of the UR3 station and the next station. The virtual model acted as a blueprint when the creation of the original kitting box was taking place..
Figure 13, Virtual model of the kitting box
With the help of the virtual model, the original kitting box could be created, which is visualized in figure 14. The kitting box consists of a pallet which is in the form of a wooden plank, that has the dimensions of 470*315 mm, located on the conveyor and is painted grey for esthetical purpose. On the pallet, there are seven different stands, that almost has the same dimensions as the stands for the shelf. The pallet also has two different metal widgets, drilled on each side of the pallet, in order for the sensors of the conveyor to recognize when the pallet arrives at one of the different stations.
Figure 14, The pallet
7.1.3 Flow-chart of the robot programming
In order to achieve a rigid program of the robot, a flow-chart was created. With the help of the flow- chart, a broad overview of what kind and in what order the robot shall pick and place the different objects. This was the independence of what was the simplest way to place widgets without colliding into other already placed widgets. The flow-chart is illustrated in figure 14. The programming of the robot consisted of spiting the flow-chart into separate programs. For instance, one program was devel- oped for only the robot to pick up and deliver sensor one, and another program was developed for picking up and deliver sensor two et cetera. The programs were then assembled into one whole pro- gram by implement-ing a sub-program that consisted of sensor 1, sensor two, et cetera. In order for the system to work re-verse, another flow-chart was created, see figure 15. The previous flow-chart was used in order to be time-efficient. However, some waypoints needed to be manipulated.
Figure 15, Flow-chart of the whole program
Figure 16, flow-chart considering reverse mode
The robot is also moving in more than 180 degrees by picking objects from the shelf and deliver to the kitting box. By reducing the risk of colliding into people in general, the robot will deliver objects by rotating to the left, seen from figure 16 perspective, see figure 16.
Figure 17, Station
7.2 Experiments
The experiments conducted was a proving point of the anthropomorphic tool regarding what kinds of objects the tool can grab. The experiments consisted of 15 different objects, both in relevance to the project in whole and natural everyday objects.
The conducted experiment is illustrated in figure 17. The grasping diameter of the object is represented as the diameter regarding the tool. The angle of the tool is measured based on the z-axis. Depending on how the tool is mounted on the robot, a different angle of the tool is changed in the form of positive and negative values, in this case, most of the angles are in negative values. The grasping percent are in the proportion of what kinds of grasping techniques being used. According to the QB Softhand research guide (2020), the maximum force of an object in a pinch grasp configuration is 11 Newton, and in power grasp, it is 62 Newton. For instance, in this case grasping Sensor one, with a pinch grasp
technique and a grasping force of 70 percent, thereby, the tool was grasping with 70 percent of 11 Newton. The differ-ent grasping techniques is visualized in figure 18.
Whenever an object is in a hanging grasp position, the maximum Newton is not described, the only constraint is the maximum payload, which is 30 kilograms. Some of the different objects, are investi- gated and is visualized in appendix B, C and D. Restriction of the different objects has been limited in the form of its weight, thereby ensuring that the tool can grab and lift the objects and not be constraint by the maximum payload of the robot.
Figure 18, Experiments conducted
Figure 19, Grasping techniques QB Softhand user guide (2020)
8 Discussion
By implementing a tool that is anthropomorphic, a concept that has been proven is to increase flexi- bility by using the tool to pick up different geometrical objects. In today's industry, struggle regarding flexibility is common, and whenever a change of product is taking place, manufacturing lines with robots' needs, in most cases, change the tool of the robot. With the help of the 19 degrees of freedom of an anthropomorphic tool, flexibility could easily be achieved.
Regarding the tool as a whole, it is limited for only using one servomotor for closing and opening, which means that whenever one wants to perform a pinch grasp, all the other fingers close. This type of configuration could easily be prevented by rotating the whole tool into a satisfying position. How- ever, it would be more successful if there were servomotors for every finger, then it would acquire flexibility like no other tool.
If, for instance, the tool would acquire a servomotor for every finger and thereby increasing even the flexibility further, one needs to consider the uncanny valley concept. If this type of configuration would be available in the future, the question needs to ask, if it is worth to produce this type of tool that has an exact resemble and configuration like a human hand to be implemented in industry. The uncanny valley concept clearly states that this kind of phobia of human alike robots or tools has a negative mental impact on the worker of an industry. The tool needs to be a selling concept for companies in the future. However, if more people or workers would be affected by the uncanny valley, the more people will be uncomfortable at the workplace.
The experiment that was conducted with 15 different objects was a success. Different grasping tech- niques have been applied to the objects to prove that it is possible to pick and place with different tool configurations. Even though if the robot would be in a tight position, were the joint becomes limited, the tool could easily be manipulated so that it could pick up and place objects in space. For picking those different objects, the outcome was expected, by using an anthropomorphic tool, references could be performed to a human hand, this was though a problem.
A lot of configurations and manipulations with the tool was performed in order to achieve a safe and firm grip of the objects. A comparison was performed between a human hand was the question was stated: "how would a human pick this object?" Whenever that was performed, programming of the robotic tool was performed where it would mimic how a human would pick the object. This did not
