- Email: [email protected]
Development of a new Mechanic Safety Coupling for Human Robot Collaboration using Magnetorheological Fluids
Available online at www.sciencedirect.com Available online at www.sciencedirect.com
ScienceDirect ScienceDirect Available online atonline www.sciencedirect.com Procedia CIRP 00 (2019) 000–000 Available at www.sciencedirect.com
ScienceDirect ScienceDirect
Procedia CIRP 00 (2019) 000–000
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Procedia CIRP 00 (2017) 000–000 Procedia CIRP 81 (2019) 908–913
52nd CIRP Conference on Manufacturing Systems 52nd CIRP Conference on Manufacturing Systems
www.elsevier.com/locate/procedia
Development of a new Mechanic Safety Coupling for Human Robot Development of aCIRP newDesign Mechanic Safety Coupling for Human Robot 28th Conference, May 2018, Nantes, France Collaboration using Magnetorheological Fluids Collaboration using Magnetorheological Fluids Lämmle* A new methodology to analyze Arik the functional and physical architecture of Arikand Lämmle* Fraunhofer Institutefor for Manufacturing Engineering Automation IPA,product Nobelstraße 12, family 70569 Stuttgart, Germany existing products an assembly oriented identification
b Second affiliation, Address, City and Postcode, Country Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Nobelstraße 12, 70569 Stuttgart, Germany b * Corresponding author. Tel.: +49-711-970-1639; fax:Second +49-711-970-1008. E-mailCity address: [email protected] affiliation, Address, and Postcode, Country
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
* Corresponding author. Tel.: +49-711-970-1639; fax: +49-711-970-1008. E-mail address: [email protected] École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
Abstract
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Abstract To ensure safety during Human-Robot-Collaboration with conventional robots, safety methods for tools are required, as common approaches don’t deliver sufficient performance. Since standard industrial robots are capable of moving heavyweight objects with high velocity and the To ensure safety during Human-Robot-Collaboration with conventional robots, safety methods for tools are required, as common approaches mechanical structure of the robot implies high moving masses, they represent a significant hazard for human operators. This paper presents an Abstract don’t deliver sufficient performance. Since standard industrial robots are capable of moving heavyweight objects with high velocity and the approach to limit the risk of injuries by reducing the transmitted energy in case of a collision by a decoupling of masses between robot and tool. mechanical structure of the robot implies high moving masses, they represent a significant hazard for human operators. This paper presents an The described tool concept is based on magnetorheological fluids, which act like an inherent safety clutch medium. Inapproach today’s business environment, the by trend towards productenergy varietyinand Dueof tomasses this development, theand needtool. of to limit the risk of injuries reducing themore transmitted casecustomization of a collisionisbyunbroken. a decoupling between robot agile and reconfigurable production emerged to cope withwhich various and product To design and optimize production The described tool concept is based systems on magnetorheological fluids, actproducts like an inherent safetyfamilies. clutch medium. © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license systems as well as to choose the optimal product matches, product analysis methods are needed. Indeed, most of the known methods aim to © 2019 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/3.0/) analyze product or one product family physical Different families, may differ largely in terms of the number and © 2019 The Authors. Published by Elsevier Ltd. This islevel. an open accessproduct article under the however, CC BY-NC-ND license This is aan open access article under the on CCthe BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems. nature of components. This fact impedes an efficient comparison and choice of appropriate product familySystems. combinations for the production (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing system. A newunder methodology is proposed to analyze existing products in view ofConference their functional and physical Systems. architecture. The aim is to cluster Peer-review responsibility of the scientific committee of the 52nd CIRP on Manufacturing Keywords: Human-Robot-Collaboration, Safety, Magnetorheological Fluids these products in new assembly oriented product families for the optimization of existing assembly lines and the creation of future reconfigurable assembly systems. Based on Datum Flow Chain, the physical structure Keywords: Human-Robot-Collaboration, Safety, Magnetorheological Fluids of the products is analyzed. Functional subassemblies are identified, and a functional analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the similarity between product families by providing design support to both, production system planners and product designers. An illustrative 1. INTRODUCTION experience during HRC applications have already been example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of developed [3]. With HRC the introduction of HRC-applications in 1. INTRODUCTION experience during applications have already been thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach. Looking at the current trend, manufacturing companies need industrial environments, the safety during needed contact in © 2017 The Authors. Published by Elsevier B.V. developed [3]. With the introduction of HRC-applications in flexibility under to frequently introduce and individualize new process or 2018. in thethe event of unintentional contact must Peer-review of the scientific committee of the 28th CIRP Designexecution Conference Looking at theresponsibility current trend, manufacturing companies need industrial environments, safety during needed contact in
products on one side while handling fluctuating customer be ensured at all times focus, a lot of contact researchmust has flexibility to frequently introduce and individualize new process execution or in[4]. theWith eventthis of unintentional – mostly gone into the characterization and development of different products on one side while handling fluctuating customer be ensured at all times [4]. With this focus, a lot of research has manually executed – reconfiguration and changeover of safety strategies, such as robotand stopping functions, demands on the other side. However, the necessary – mostly gone into the characterization development of collision different production lines are not consistent with the current automation detection or limiting the motion of the robot [5]. Furthermore, manually executed – reconfiguration and changeover of safety strategies, such as robot stopping functions, collision paradigm of strict spatial separation of the human worker and various safety measures have been developed in Furthermore, recent years production lines are not consistent with the current automation detection or limiting the of the robot [5]. 1.the Introduction of the product range andmotion characteristics manufactured and/or robot. Thereby superior cognitive and sensory skills of the pursuing the aim to making separating safety fences obsolete. paradigm of strict spatial separation of the human worker and various safety measures have been developed in recent years assembled in this system. In this context, the main challenge in human worker are neither used during process execution nor Collaborative robots in separating industrial safety robot fences applications are theDue robot.toThereby superior cognitive and sensory skills of the pursuing the aim to making obsolete. the fast development in the domain of modelling and analysis is now not only to cope with single combined with the robot’s capabilities of handling heavy usually small – robots so-called sensitive lightweight robots (or human worker are neither used during execution and nor Collaborative inrange industrial robot product applications are communication an ongoing trendprocess of digitization products, a and limited product orinexisting families, objects with highand persistence and accuracy. cobots) – are implemented a defined collaboration combined with the robot’s capabilities of handling heavy usually small – so-called sensitive lightweight robots (or digitalization, manufacturing enterprises are facing important but also to be able to analyze and tofor compare products the to define workspace with integrated sensors safely stopping robot objects with high persistence and accuracy. cobots) – and are implemented in a defined collaboration Flexible Human Robot Collaboration (HRC) offers a challenges in today’s market environments: a continuing new product families. It can be observed that classical existing after a collision is detected. These lightweight robotsthe onrobot one workspace with are integrated sensors for safely stopping possible towards contribution to satisfy nowadays needs in production tendency reduction of product development times and product families regrouped in function of clients or features. Flexible Human Robot Collaboration (HRC) offers a side show a substantially reduced risk of possible injuries for after a collision is oriented detected.product These lightweight onfind. one lines within the lifecycles. field of tension between “Flexibility” and shortened product In addition, is an increasing However, assembly families are robots hardly to possible contribution to satisfy nowadaysthere needs in production the human worker as a result of a collision during HRC, due to side a substantially reduced risk of differ possible injuries for “Efficiency”. Several methods to task demand of customization, at improve the samethe timeefficient in a global Onshow the moved product family On level, products mainly in two lines within the field of being tension between “Flexibility” and their low masses. the other side, lightweight robots the human worker as(i) a result of a collision during HRC, duethe to allocation – with by assigning process tasksthe to the human competition competitors Thisworker trend, main characteristics: the number of andclassical (ii) “Efficiency”. Several methodsalltoover improveworld. the efficient task forfeit some of their characteristics in components comparison to their low moved masses. On the other side, lightweight robots and the robot by best suitability [1, 2] – as well as the user which is inducing the development macro to worker micro type of components (e.g. mechanical, electrical, allocation – by assigning process tasksfrom to the human industrial robots such as speed, payload, stiffnesselectronical). andtoextensive forfeit some of their characteristics in comparison classical markets, results in diminished lot sizes due to augmenting Classical methodologies considering mainly single products and the robot by best suitability [1, 2] – as well as the user industrial robots such as speed, payload, stiffness and extensive product varieties (high-volume to low-volume production) [1]. or solitary, already existing product families analyze the 2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which (http://creativecommons.org/licenses/by-nc-nd/3.0/) 2212-8271possible ©under 2019responsibility The optimization Authors. of Published by Elsevier Ltd. This an open access article under the CC BY-NC-ND licensean efficient definition and identify potentials in of the existing causes regarding Peer-review the scientific committee theis52nd CIRP Conference on difficulties Manufacturing Systems. (http://creativecommons.org/licenses/by-nc-nd/3.0/) production system, it is important to have a precise knowledge comparison of different product families. Addressing this demandsAssembly; on the other However, the necessary Keywords: Designside. method; Family identification
Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems.
2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an©open article Published under theby CC BY-NC-ND 2212-8271 2017access The Authors. Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of scientific the scientific committee theCIRP 52ndDesign CIRPConference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2019.03.226
2
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
movement in favor of inherent safety. However, these characteristics promise feasible execution of specific processes concerning achievable cycle time and therefore economic efficiency. To enable classical robots for safe HRC new safety measures are needed. In this paper, a new mechanical coupling device using magnetorheological fluids (MRF) for safe human robot collaboration is presented. It aims at reducing the transmitted kinetic energy into the human body during a collision with a robot manipulator. This is goal is realized through a mass decoupling device mounted between the robot flange and the tool. The paper is structured as follows: In section 2 related work in the field of safety measures for HRC as well as magnetorheological fluids are explored. Section 3 focusses on the operating principle and the design of the mechanic coupling device. The development regarding the magnetic field design and activation of the MRF are described in section 4. Subsequently validation and experimental investigation of the coupling device are discussed in section 5. Finally, Section 6 sums up the conducted steps and presents an outlook on future research and development activities. 2. State of the Art Many novel sensors – partially even applicable to classical industrial robots – as well as new robot kinematics have been developed in recent years to realize safe HRC applications. A short overview of the current state of research and state of the art is given in the following. 2.1. Novel robot kinematics and safety sensors for HRC As already mentioned beforehand, robots applied in human robot collaboration are almost exclusively safe, lightweight kinematics, using active control, passive compliance or a combination of both concepts. [6] Furthermore, many new approaches focus on the definition of static and dynamic safety zones around the robot. Novel safety sensors track the violation of these zones to prevent collisions between the human worker and the robot system, e.g. using vision systems [7, 8], light or laser detection and ranging (lidar/ladar) [9] as well as tactile sensors in the floor. [10, 11] A lot of research has also gone into the development of new robot skins to detect the human worker before a collision or at contact. Several of these skins aim to reduce the applied forces and the impact occurring during a collision. For this purpose, Kim et. al. developed a “soft skin module with a built-in airtight cavity in which air pressure can be sensed” [12]. A similar approach is used by Blue Danube Robotics (Austria) for the AirSkin. [13] In general, a wide variety of different sensing technologies are used in robot skins to detect the human worker. Faude (Germany) as well as Bosch (Germany) provide already commercially available solutions using tactile skins for collision detection and protection. Other approaches to be mentioned are the Hex-o-Skin developed by the Technical University of Munich [14, 15] and the “sensorized flexible skin” presented by Cirillo et. al. in [16].
909
2.2. Force and injury thresholds Beside novel approaches in robot kinematics as well as a wide variety of safety measures to avoid contact between a human worker and the robot, a lot of research has gone into determing thresholds for collisions during HRC-applications. Oberer-Treitz et. al. defines the safety-paradigm of “Passive Safety” in [17] as all safety measures “to reduce the effects in case of a collision, in contrast to those active safety measure that try to avoid the collision.” Kokkalis et. al. presented an approach for detecting external forces using an inverse dynamic model of the robot [18]. Haddadin et. al. evaluated differences concerning non-constrained and constrained blunt impacts as well as the role of the robot mass and velocity, leding “to the conclusion that no robot whatever mass it has can become life threatening at typical robot speeds if the human is not clamped” [19]. Based on the DLR Crash Report [20] as well as several other investigations, the ISO/TS 15066 was drafted and further refined. ISO/TS 15066 provides necessary force and pressure thresholds for quasi-static (clamping) and transient collision as well as limits for transmitted energy depending on the affected region of the human body [21]. All defined thresholds were evaluated within experimental medical studies by Behrens and Elkmann in [22]. Hong et. al. analyzed the effect of impact-absorbing depending on the joints type and described safe configurations for serial-chain manipulators [23]. 2.3. Magnetorheological fluids Magnetorheological fluids (MRF) are so called “smart materials”. Applying an external magnetic field and directing it through the MRF changes the mechanical characteristics of the fluid within only a few milliseconds, from a low viscosity to an approximately solid state of the fluid. This behavior offers the possibility to lock as well as release the movement of the mechanic safety coupling. Since the amount of transmittable force through the MRF is dependent on the characteristics of the magnetic field, the fluid itself functions as an inherently safe coupling substance. Limiting the transmittable force to a safety threshold defined in ISO/TS 15066 by changing the magnetic field, enables a movement of the coupling before the collision is detected. MRF are used for example in technical applications such as mechanical dampers, clutches and breaks, utilizing their ability of a characteristic’s change within a very short period of time. [24, 25]Shafer and Kermani present in [26]and [27]the design of a two degrees of freedom safe manipulator with a single motor. The manipulator/robot uses magnetorheological fluids within its clutches to control the transmitted torque and therefore enable safe usage within human robot collaboration. Each joint of the robot combines two MRF-clutches, which shut the movement in opposite directions. The movement of each joint is controlled using the magnetic field and changing the behavior of the MRF within each clutch. The safety of the robot during HRC applications is ensured by limiting the transmittable torque by each of the robot’s joints.
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle/ Procedia CIRP 00 (2019) 000–000
910
3. Mechanic Safety Coupling Design Using the mechanical coupling unit reduces the transmitted kinetic energy from the robot system on the human worker in case of a collision, as already mentioned in Section 1. The human body is not capable of absorbing the high, transmitted kinetic energy, especially during a collision with a classical industrial robot, without potentially suffering serious injuries. Therefore, ISO/TS 15066 defines the acceptable limits for the transmitted energy based on the biomechanical threshold values for each region of the human body. Furthermore, the standard defines the calculation of the maximum relative velocity between the human worker and the robot. Since quasi-static contacts represent a significant higher hazard potential, designing a robot application with respect to quasi-static as well as transient collisions limits the maximum robot velocity and therefore the economic efficiency of the application. Following interrelationships are only valid for the transient contact. It is based on the assumption of the worstcase scenario, a complete inelastic contact situation, as shown in Figure 1. Here, the relative kinetic energy is being completely transferred to the affected body part.
Fig. 1. Contact model of a complete inelastic contact situation during a transient collision with mR being the effective mass of the robot depending on the robot position and movement, mH being the effective mass of the affected body part of the human worker, A being the area of contact between the robot and the part of the human body and vrel being the relative velocity between the robot and the affected body part.
The transmitted kinetic energy equals the change of the intrinsic energy as a result of the collision. Exceedance of the defined threshold values for the energy transmission is prohibited and must be taken into account during design of the HRC application. This value takes the maximum contact force Fmax as well as the stiffness k, depending on the affected part of the human body, into account: 𝐸𝐸 =
2 𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚
2𝑘𝑘
=
2 𝐴𝐴2 𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚
3
Specific values for mH for the application of equation (4) are listed in ISO/TS 15066.
Fig. 2. Simplified model of the mass distribution on a robot [17].
The effective mass of the robot mR depends on its actual position and movement, captured in the total mass M of all moved parts of the system, as well as the effective payload mL including the tool and the workpiece: 𝑚𝑚𝑅𝑅 =
𝑀𝑀 2
+ 𝑚𝑚𝐿𝐿
(5)
Heavy, classical industrial robots, in comparison to lightweight robots, are not slowed down significantly by the applied contact forces. Equations (2) to (4) show the possibility of increasing the maximum relative velocity of the robot during HRC applications by substantially reducing the effective mass of the robot, while respecting the maximum contact forces. Therefore, decoupling the heavy mass of the classical industrial robot from the effective mass of the tool as well as possible handling objects, during the braking process of the robot after a collision has been detected, shows a suitable approach for enabling usage of these robots during human robot collaboration. By decoupling the mass and enabling an almost complete free translational movement of the remaining collision mass (Fig. 3 a), only transient contact situations occur since a clamping of the human is not possible (Fig. 3 b).
(1)
2𝑘𝑘
Using the area of contact A between the robot and the body part, the maximum contact force Fmax can also be translated to a maximum contact pressure pmax. Comparing equation (1) with the complete kinetic energy at the center of gravity of the mass can be used to calculate the maximum acceptable relative velocity vrel: 𝐸𝐸 =
𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚 2 2𝑘𝑘
𝑣𝑣𝑟𝑟𝑟𝑟𝑟𝑟,𝑚𝑚𝑚𝑚𝑚𝑚 =
1
2 = 𝜇𝜇𝑚𝑚 𝑣𝑣𝑟𝑟𝑟𝑟𝑟𝑟,𝑚𝑚𝑚𝑚𝑚𝑚 2
𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚
√𝜇𝜇𝑚𝑚𝑘𝑘
=
𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚 𝐴𝐴 √𝜇𝜇𝑚𝑚 𝑘𝑘
(2) (3)
Here µm captures the mass of the reduced two body system, consisting of the effective mass of the robot mR as well as the effective mass mH of the affected part of the human body, as also shown in Figure 2: 𝜇𝜇𝑚𝑚 = (
1
𝑚𝑚𝐻𝐻
+
1
𝑚𝑚𝑅𝑅
)
−1
(4)
Fig. 3. Mechanical safety coupling (a) with two translational degrees of freedom. Defined by the parallel kinematic of the coupling, plane E2 and plane E3 are always oriented parallel towards each other. The coupling can be mounted on the robot flange using the mechanical interface 2. The tool is connected using the mechanical interface 3.; (b) In case of a collision with the human worker (P), the robot’s (R) mass is decoupled from the tool and workpiece mass (W) which therefore prevents collisions with the full mass of the robotic system as well as quasi-static contacts at all.
4
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
For decoupling the mass of the robot from the mass of the tool within a very short time, specific actuators are needed. For this purpose, magnetorheological fluids (MRF) were identified as an inherent safe substance for actuating the mechanic coupling. The transmittable force and therefore the maximum load of the coupling is defined by the surface area covered with MRF. So far, the coupling is designed for a maximum payload of 10 kg in all translational degrees of freedom. Increasing the area covered with MRF also increases the maximum load of the coupling. The mechanic coupling itself is designed with three pairs of MRF actuators. These actuators function as ball joints and are connected with rigid bars (Fig. 4 a). This design also allows By varying the length of the bars, the braking distance of different robots can be compensated. Each ball joint is filled with magnetorheological fluid surrounding the ball itself to lock as well as release the movement of the joint. Since MRF are activated, i.e. change from a low viscosity fluid to an approximately solid state, by magnetic fields, the joint also contains several magnetic field cores. The cross section of a ball joint is given in Figure 4 b).
911
the magnetorheological fluid. Energizing each coil around a core separately and without a magnetic connection to other cores leads to misdirection and vortexes of the magnetic field. Since therefore a directed movement of the magnetic field lines through the MRF is not possible, a sufficient activation of the fluid doesn not take place. 4.2. Magnetic field bridges To overcome this problem of insufficient penetration of the MRF with the magnetic field, the magnetic cores next to each other are connected using a magnetic field bridge to create closed magnetic circuits (Fig. 6). Therefore, the complete magnetic flux is focused and directed through the MRF.
Fig. 5. Closed magnetic circuit within the ball joint, with the magnetic resistance of the MRF ℜ𝑀𝑀𝑀𝑀𝑀𝑀 , of the steel ball ℜ𝑠𝑠1 and the magnetic bridge ℜ𝑠𝑠2 , as well as the magnetic resistances of the cores ℜ𝑐𝑐1 , ℜ𝑐𝑐2 and the threaded rods ℜ𝑠𝑠𝑠𝑠1 and ℜ𝑠𝑠𝑠𝑠2 .
Fig. 4. Cross section of the ball joints of the mechanical safety coupling. (a) The coupling is designed using three pairs of ball joints each connected with a rigid bar (4). The ball itself (7) is surrounded by MRF (11). Balls in comparison to other geometries offer a bigger surface area (9) to be covered by the MRF which increases the maximum assignable force. Sealing rings (10) are used to prevent leakage of the fluid during regular usage. The magnetic field is generated with multiple magnetic field cores (12). (b) Enlarged display of the cross section of a ball joint.
The magnetorheological fluid itself should represent the highest magnetic resistance ℜ𝑀𝑀𝑀𝑀𝑀𝑀 within the circuit. As a result, the highest magnetic potential difference is present at the MRF. Simulative examination of the adjusted magnetic field design is shown in Figure 6.
4. Magnetic Field Design Besides the mechanical design of the coupling, the technical configuration of the magnetic fields is of superior importance for the functionality of the safety coupling. For this reason, simulation studies using ANSYS Workbench were carried out. The design of the coupling focusing on the magnetic design is described in the following. 4.1. Magnetic field cores and coils Overall, seven magnetic field cores are positioned around the ball to achieve a comprehensive penetration of the MRF with the magnetic field. Each magnetic field core is produced out of soft magnetic cobalt-iron alloy and wrapped with multiple layers of enamelled copper wire. The cores are mounted to the housing of the ball joint using threaded rods and are being oriented to realize a directed magnetic flux through
Fig. 6. Simulation of the magnetic flux density in the ball joint with adjusted magnetic field design. Since threaded rods are used to assemble the cores, the air within the threaded holes cut in the cores represents an area of high loss of the magnetic flux within simulation. This problem was fixed screwing the threaded rods completely into the core.
5. Experimental Studies and Validation To analyze the behavior of the coupling and the potential for integration in a robotic application, several collision experiments were performed. Goal of the experiments was to investigate the applied contact forces during a collision as well as the contact situation of the collision. The experimental setup and the results will be discussed in the following.
912
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle/ Procedia CIRP 00 (2019) 000–000
5
5.1. Experimental setup and collision procedure Two ball joints connected with a rod are used for the collision experiments. The upper ball joint is mounted on a UR5 lightweight robot. An acceleration sensor is used to detect a collision. All applied forces during the collision are recorded using a force measuring device especially designed for force examination within HRC applications. The complete experimental setup is shown in Figure 7. Beside the contact forces, the velocity of the robot during the collision experiments is recorded as well.
Fig. 7. Experimental setup for collision experiments using the mechanic safety coupling and a UR5 industrial robot by Universal Robots.
During the collision experiments, the robot first approaches a position in front of the force measuring device. Subsequently the robot accelerates until it has reached the previously defined velocity and collides with the force measuring device. In the process of a collision, the acceleration sensor sends a signal to the coupling control to deactivate the MRF as well as to the robot controller, which then stops the robot. The collision experiments are carried out under variation of the robot’s velocity. So far, the reaction time of the acceleration sensor and the time needed to decouple the mass of the robot from the mass of the tool by deactivation the MRF was not evaluated in the experiments, but will be part of future work.
Fig. 8. Force progression during collision for a robot velocity of vIR=900mm/s with multiple characteristic force impacts with decreasing intensity.
As already mentioned by Haddadin et. al. in [15], the applied forces during collision are highly dependent on the velocity of the robot. Therefore increasing the velocity of the robot leads to higher forces measured during the collision experiments. Figure 9 displays the maximum applied forces with standard deviation for different velocities of the robot. The increasing variance of the standard deviation is due to the manual positioning of the ball joint at the beginning of each collision experiment. Already minor torsion of the lower ball joint results in variable contact geometries and therefore in differences concerning the maximum applied forces. Overall the maximum measured forces during the collision experiments 𝐹𝐹𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 = 213 𝑁𝑁 did not exceed the lowest allowed force threshold of 𝐹𝐹𝑇𝑇,𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠ℎ = 220𝑁𝑁 (excluding the critical body parts face, forehead and head; here, collisions should be strictly avoided) defined by ISO/TS 15066, even for a robot velocity of 𝑣𝑣𝐼𝐼𝐼𝐼 = 900 𝑚𝑚𝑚𝑚⁄𝑠𝑠 .
5.2. Experimental results The force progressions – showed exemplary in Figure 8 – over all experiments indicate an exclusively transient contact situation under usage of the mechanical safety coupling. At the beginning of each force progression, the impact of the ball joint at the force measuring device is indicated by a single force pulse of a few milliseconds, followed by several short impacts with reduced force. This decreasing intensity of the applied force indicates a transmission of kinetic energy from the moved system on to the fixed force measuring device. Additional force impacts after the initial one are only observable from a robot velocity of 𝑣𝑣𝐼𝐼𝐼𝐼 = 250 𝑚𝑚𝑚𝑚⁄𝑠𝑠 upwards. This circumstance is explained by the mechanical design of the force measurement device. Below the mentioned velocity threshold, the force impact is too minor to compress the integrated spring enough to push away the ball joint and cause further impacts.
Fig. 9. Average maximum force with standard variation for all conducted collision experiments under variation of the robot velocity.
Using the measured maximum force during collision, an estimation of applicable robot velocities with integration of the mechanical safety coupling is possible. For this purpose, the measured forces can be matched with the maximum contact forces defined in ISO/TS 15066 for potentially affected body parts. The initial design of the coupling was developed to carry out first collision experiments. For this purpose, the coupling itself was positioned manually after each collision. Future designs will also have a mechanism to lock the coupling in an initial position to ensure accuracy.
6
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
6. Discussion and Outlook In this paper, the development of a new mechanical safety coupling using magnetorheological fluids was evaluated. It was shown, that decoupling the effective mass of the robot from the effective mass of the tool as well as potentially handled parts during collision, significantly reduces the transmitted energy upon the affected parts of the human body. For this purpose, the MRF acts as an inherently safe coupling substance, since high external forces. e.g. as a result of a collision, already lead to a deformation of the fluid and therefore to a decoupling of the masses. The application of the mechanical safety coupling under real-world conditions was evaluated. The collision experiments point out an exclusively transient contact situation under usage of the mechanical safety coupling. This circumstance allows the design of the application with respect to the higher biomechanical thresholds for transient contact situations compared to quasi-static contacts defined in the standards. Furthermore, the robot is stopped solely by its own brakes and because of the decoupling of the robot mass and the mass of the tool not additionally through transmitting kinetic energy upon the collision partner. Using the safety coupling, the measured contact forces during the collision experiments did not exceed the defined force thresholds even for robot velocities beyond typical velocities of 500 mm/s in HRCapplications. Optimizing the magnetic field design, especially directing the movement of the magnetic field lines through the MRF and increasing the magnetic flux density within the MRF will be the focus of the author’s ongoing work. Furthermore, advanced collision experiments using classical industrial robots will be performed in the future. For this purpose a mechanically and magnetically optimized safety coupling, with increased payload will be developed. Since magnetorheological fluids tend to separate their individual ingredients, these experiments will also cover a detailed analysis of the long-term behavior of the mechanic safety coupling. References [1] Michalos, G., Makris, S., Tsarouchi, P., Guasch, T. et al., 2015. Design Considerations for Safe Human-robot Collaborative Workplaces 37, p. 248. [2] Santis, A. de, Siciliano, B., Luca, A. de, Bicchi, A. An atlas of physical human–robot interaction, in Mechanism and Machine Theory vol. 43, p. 253. [3] Lenz, C., Sotzek, A., Roder, T., Radrich, H. et al., 2011. Human workflow analysis using 3D occupancy grid hand tracking in a human-robot collaboration scenario, in 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, IEEE, p. 3375.
913
[4] Basso, F., Munaro, M., Michieletto, S., Pagello, E. et al., 2013. Fast and Robust Multi-people Tracking from RGB-D Data for a Mobile Robot, in Intelligent Autonomous Systems 12, Springer Berlin Heidelberg, Berlin, Heidelberg, p. 265. [5] Taipalus, T., Ahtiainen, J., 2011. Human detection and tracking with kneehigh mobile 2D LIDAR, in 2011 IEEE International Conference on Robotics and Biomimetics, IEEE, p. 1672. [6] Rio, M., Colas, F., Andries, M., Charpillet, F., 2016. Probabilistic sensor data processing for robot localization on load-sensing floors, in 2016 IEEE International Conference on Robotics and Automation (ICRA), IEEE, p. 4544. [7] Vogel, C., Fritzsche, M., Elkmann, N., 2016. Safe human-robot cooperation with high-payload robots in industrial applications, in 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI), IEEE, p. 529. [8] Kim, J., Alspach, A., Yamane, K., 2015. 3D printed soft skin for safe human-robot interaction, in 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, p. 2419. [9] Blue danube robotics. Technology for robots. https://www.bluedanuberobotics.com/airskin/. [10] Mittendorfer, P., Cheng, G. Uniform Cellular Design of Artificial Robotic Skin, in 2012 7th German Conference on Robotics, p. 1. [11] Mittendorfer, P., Cheng, G. Humanoid Multimodal Tactile-Sensing Modules, in IEEE Transactions on Robotics, vol. 27, no 3, 2011, p. 401. [12] Cirillo, A., Ficuciello, F., Natale, C., Pirozzi, S. et al. A Conformable Force/Tactile Skin for Physical Human–Robot Interaction, in 2016 IEEE Robotics and Automation Letters, vol. 1, no.1, p. 41. [13] Oberer-Treitz, S., Dietz, T., Verl, A., 2013. Safety in industrial applications: From fixed fences to direct interaction, in IEEE ISR 2013, IEEE, p. 1. [14] Haddadin, S., Albu-Schaffer, A., Frommberger, M., Hirzinger, G., 2008. The role of the robot mass and velocity in physical human-robot interaction - Part II: Constrained blunt impacts, in 2008 IEEE International Conference on Robotics and Automation, IEEE, p. 1339. [15] Haddadin, S., Albu-Schaffer, A., Frommberger, M., Rossmann, J. et al., 2009. The “DLR Crash Report”: Towards a standard crash-testing protocol for robot safety - Part I: Results, in 2009 IEEE International Conference on Robotics and Automation, IEEE, p. 272. [16] Deutsches Institut für Normung. DIN ISO/TS 15066 Roboter und Robotikgeräte - Kollaborierende Roboter. [17] Behrens, R., Elkmann, N., 2014. Study on meaningful and verified Thresholds for minimizing the consequences of human-robot collisions, in 2014 IEEE International Conference on Robotics and Automation (ICRA), IEEE, p. 3378. [18] Spaggiari, A. Properties and applications of Magnetorheological fluids, in 2012 Frattura ed Integrità Strutturale, 23, p. 48. [19] Wang, J., Meng, G. Magnetorheological fluid devices: Principles, characteristics and applications in mechanical engineering, in Proceedings of the IMechE, p. 165. [20] Yadmellat, P., Shafer, A.S., Kermani, M.R. Design and Development of a Single-Motor, Two-DOF, Safe Manipulator, in IEEE/ASME Transactions on Mechatronics, p. 1384. [21] Shafer, A.S., Kermani, M.R., 2009. On the feasibility and suitability of MR and ER based actuators in human friendly manipulators, in 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, IEEE, p. 2904.
ScienceDirect ScienceDirect Available online atonline www.sciencedirect.com Procedia CIRP 00 (2019) 000–000 Available at www.sciencedirect.com
ScienceDirect ScienceDirect
Procedia CIRP 00 (2019) 000–000
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Procedia CIRP 00 (2017) 000–000 Procedia CIRP 81 (2019) 908–913
52nd CIRP Conference on Manufacturing Systems 52nd CIRP Conference on Manufacturing Systems
www.elsevier.com/locate/procedia
Development of a new Mechanic Safety Coupling for Human Robot Development of aCIRP newDesign Mechanic Safety Coupling for Human Robot 28th Conference, May 2018, Nantes, France Collaboration using Magnetorheological Fluids Collaboration using Magnetorheological Fluids Lämmle* A new methodology to analyze Arik the functional and physical architecture of Arikand Lämmle* Fraunhofer Institutefor for Manufacturing Engineering Automation IPA,product Nobelstraße 12, family 70569 Stuttgart, Germany existing products an assembly oriented identification
b Second affiliation, Address, City and Postcode, Country Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Nobelstraße 12, 70569 Stuttgart, Germany b * Corresponding author. Tel.: +49-711-970-1639; fax:Second +49-711-970-1008. E-mailCity address: [email protected] affiliation, Address, and Postcode, Country
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
* Corresponding author. Tel.: +49-711-970-1639; fax: +49-711-970-1008. E-mail address: [email protected] École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
Abstract
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Abstract To ensure safety during Human-Robot-Collaboration with conventional robots, safety methods for tools are required, as common approaches don’t deliver sufficient performance. Since standard industrial robots are capable of moving heavyweight objects with high velocity and the To ensure safety during Human-Robot-Collaboration with conventional robots, safety methods for tools are required, as common approaches mechanical structure of the robot implies high moving masses, they represent a significant hazard for human operators. This paper presents an Abstract don’t deliver sufficient performance. Since standard industrial robots are capable of moving heavyweight objects with high velocity and the approach to limit the risk of injuries by reducing the transmitted energy in case of a collision by a decoupling of masses between robot and tool. mechanical structure of the robot implies high moving masses, they represent a significant hazard for human operators. This paper presents an The described tool concept is based on magnetorheological fluids, which act like an inherent safety clutch medium. Inapproach today’s business environment, the by trend towards productenergy varietyinand Dueof tomasses this development, theand needtool. of to limit the risk of injuries reducing themore transmitted casecustomization of a collisionisbyunbroken. a decoupling between robot agile and reconfigurable production emerged to cope withwhich various and product To design and optimize production The described tool concept is based systems on magnetorheological fluids, actproducts like an inherent safetyfamilies. clutch medium. © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license systems as well as to choose the optimal product matches, product analysis methods are needed. Indeed, most of the known methods aim to © 2019 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/3.0/) analyze product or one product family physical Different families, may differ largely in terms of the number and © 2019 The Authors. Published by Elsevier Ltd. This islevel. an open accessproduct article under the however, CC BY-NC-ND license This is aan open access article under the on CCthe BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems. nature of components. This fact impedes an efficient comparison and choice of appropriate product familySystems. combinations for the production (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing system. A newunder methodology is proposed to analyze existing products in view ofConference their functional and physical Systems. architecture. The aim is to cluster Peer-review responsibility of the scientific committee of the 52nd CIRP on Manufacturing Keywords: Human-Robot-Collaboration, Safety, Magnetorheological Fluids these products in new assembly oriented product families for the optimization of existing assembly lines and the creation of future reconfigurable assembly systems. Based on Datum Flow Chain, the physical structure Keywords: Human-Robot-Collaboration, Safety, Magnetorheological Fluids of the products is analyzed. Functional subassemblies are identified, and a functional analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the similarity between product families by providing design support to both, production system planners and product designers. An illustrative 1. INTRODUCTION experience during HRC applications have already been example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of developed [3]. With HRC the introduction of HRC-applications in 1. INTRODUCTION experience during applications have already been thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach. Looking at the current trend, manufacturing companies need industrial environments, the safety during needed contact in © 2017 The Authors. Published by Elsevier B.V. developed [3]. With the introduction of HRC-applications in flexibility under to frequently introduce and individualize new process or 2018. in thethe event of unintentional contact must Peer-review of the scientific committee of the 28th CIRP Designexecution Conference Looking at theresponsibility current trend, manufacturing companies need industrial environments, safety during needed contact in
products on one side while handling fluctuating customer be ensured at all times focus, a lot of contact researchmust has flexibility to frequently introduce and individualize new process execution or in[4]. theWith eventthis of unintentional – mostly gone into the characterization and development of different products on one side while handling fluctuating customer be ensured at all times [4]. With this focus, a lot of research has manually executed – reconfiguration and changeover of safety strategies, such as robotand stopping functions, demands on the other side. However, the necessary – mostly gone into the characterization development of collision different production lines are not consistent with the current automation detection or limiting the motion of the robot [5]. Furthermore, manually executed – reconfiguration and changeover of safety strategies, such as robot stopping functions, collision paradigm of strict spatial separation of the human worker and various safety measures have been developed in Furthermore, recent years production lines are not consistent with the current automation detection or limiting the of the robot [5]. 1.the Introduction of the product range andmotion characteristics manufactured and/or robot. Thereby superior cognitive and sensory skills of the pursuing the aim to making separating safety fences obsolete. paradigm of strict spatial separation of the human worker and various safety measures have been developed in recent years assembled in this system. In this context, the main challenge in human worker are neither used during process execution nor Collaborative robots in separating industrial safety robot fences applications are theDue robot.toThereby superior cognitive and sensory skills of the pursuing the aim to making obsolete. the fast development in the domain of modelling and analysis is now not only to cope with single combined with the robot’s capabilities of handling heavy usually small – robots so-called sensitive lightweight robots (or human worker are neither used during execution and nor Collaborative inrange industrial robot product applications are communication an ongoing trendprocess of digitization products, a and limited product orinexisting families, objects with highand persistence and accuracy. cobots) – are implemented a defined collaboration combined with the robot’s capabilities of handling heavy usually small – so-called sensitive lightweight robots (or digitalization, manufacturing enterprises are facing important but also to be able to analyze and tofor compare products the to define workspace with integrated sensors safely stopping robot objects with high persistence and accuracy. cobots) – and are implemented in a defined collaboration Flexible Human Robot Collaboration (HRC) offers a challenges in today’s market environments: a continuing new product families. It can be observed that classical existing after a collision is detected. These lightweight robotsthe onrobot one workspace with are integrated sensors for safely stopping possible towards contribution to satisfy nowadays needs in production tendency reduction of product development times and product families regrouped in function of clients or features. Flexible Human Robot Collaboration (HRC) offers a side show a substantially reduced risk of possible injuries for after a collision is oriented detected.product These lightweight onfind. one lines within the lifecycles. field of tension between “Flexibility” and shortened product In addition, is an increasing However, assembly families are robots hardly to possible contribution to satisfy nowadaysthere needs in production the human worker as a result of a collision during HRC, due to side a substantially reduced risk of differ possible injuries for “Efficiency”. Several methods to task demand of customization, at improve the samethe timeefficient in a global Onshow the moved product family On level, products mainly in two lines within the field of being tension between “Flexibility” and their low masses. the other side, lightweight robots the human worker as(i) a result of a collision during HRC, duethe to allocation – with by assigning process tasksthe to the human competition competitors Thisworker trend, main characteristics: the number of andclassical (ii) “Efficiency”. Several methodsalltoover improveworld. the efficient task forfeit some of their characteristics in components comparison to their low moved masses. On the other side, lightweight robots and the robot by best suitability [1, 2] – as well as the user which is inducing the development macro to worker micro type of components (e.g. mechanical, electrical, allocation – by assigning process tasksfrom to the human industrial robots such as speed, payload, stiffnesselectronical). andtoextensive forfeit some of their characteristics in comparison classical markets, results in diminished lot sizes due to augmenting Classical methodologies considering mainly single products and the robot by best suitability [1, 2] – as well as the user industrial robots such as speed, payload, stiffness and extensive product varieties (high-volume to low-volume production) [1]. or solitary, already existing product families analyze the 2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which (http://creativecommons.org/licenses/by-nc-nd/3.0/) 2212-8271possible ©under 2019responsibility The optimization Authors. of Published by Elsevier Ltd. This an open access article under the CC BY-NC-ND licensean efficient definition and identify potentials in of the existing causes regarding Peer-review the scientific committee theis52nd CIRP Conference on difficulties Manufacturing Systems. (http://creativecommons.org/licenses/by-nc-nd/3.0/) production system, it is important to have a precise knowledge comparison of different product families. Addressing this demandsAssembly; on the other However, the necessary Keywords: Designside. method; Family identification
Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems.
2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an©open article Published under theby CC BY-NC-ND 2212-8271 2017access The Authors. Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of scientific the scientific committee theCIRP 52ndDesign CIRPConference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2019.03.226
2
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
movement in favor of inherent safety. However, these characteristics promise feasible execution of specific processes concerning achievable cycle time and therefore economic efficiency. To enable classical robots for safe HRC new safety measures are needed. In this paper, a new mechanical coupling device using magnetorheological fluids (MRF) for safe human robot collaboration is presented. It aims at reducing the transmitted kinetic energy into the human body during a collision with a robot manipulator. This is goal is realized through a mass decoupling device mounted between the robot flange and the tool. The paper is structured as follows: In section 2 related work in the field of safety measures for HRC as well as magnetorheological fluids are explored. Section 3 focusses on the operating principle and the design of the mechanic coupling device. The development regarding the magnetic field design and activation of the MRF are described in section 4. Subsequently validation and experimental investigation of the coupling device are discussed in section 5. Finally, Section 6 sums up the conducted steps and presents an outlook on future research and development activities. 2. State of the Art Many novel sensors – partially even applicable to classical industrial robots – as well as new robot kinematics have been developed in recent years to realize safe HRC applications. A short overview of the current state of research and state of the art is given in the following. 2.1. Novel robot kinematics and safety sensors for HRC As already mentioned beforehand, robots applied in human robot collaboration are almost exclusively safe, lightweight kinematics, using active control, passive compliance or a combination of both concepts. [6] Furthermore, many new approaches focus on the definition of static and dynamic safety zones around the robot. Novel safety sensors track the violation of these zones to prevent collisions between the human worker and the robot system, e.g. using vision systems [7, 8], light or laser detection and ranging (lidar/ladar) [9] as well as tactile sensors in the floor. [10, 11] A lot of research has also gone into the development of new robot skins to detect the human worker before a collision or at contact. Several of these skins aim to reduce the applied forces and the impact occurring during a collision. For this purpose, Kim et. al. developed a “soft skin module with a built-in airtight cavity in which air pressure can be sensed” [12]. A similar approach is used by Blue Danube Robotics (Austria) for the AirSkin. [13] In general, a wide variety of different sensing technologies are used in robot skins to detect the human worker. Faude (Germany) as well as Bosch (Germany) provide already commercially available solutions using tactile skins for collision detection and protection. Other approaches to be mentioned are the Hex-o-Skin developed by the Technical University of Munich [14, 15] and the “sensorized flexible skin” presented by Cirillo et. al. in [16].
909
2.2. Force and injury thresholds Beside novel approaches in robot kinematics as well as a wide variety of safety measures to avoid contact between a human worker and the robot, a lot of research has gone into determing thresholds for collisions during HRC-applications. Oberer-Treitz et. al. defines the safety-paradigm of “Passive Safety” in [17] as all safety measures “to reduce the effects in case of a collision, in contrast to those active safety measure that try to avoid the collision.” Kokkalis et. al. presented an approach for detecting external forces using an inverse dynamic model of the robot [18]. Haddadin et. al. evaluated differences concerning non-constrained and constrained blunt impacts as well as the role of the robot mass and velocity, leding “to the conclusion that no robot whatever mass it has can become life threatening at typical robot speeds if the human is not clamped” [19]. Based on the DLR Crash Report [20] as well as several other investigations, the ISO/TS 15066 was drafted and further refined. ISO/TS 15066 provides necessary force and pressure thresholds for quasi-static (clamping) and transient collision as well as limits for transmitted energy depending on the affected region of the human body [21]. All defined thresholds were evaluated within experimental medical studies by Behrens and Elkmann in [22]. Hong et. al. analyzed the effect of impact-absorbing depending on the joints type and described safe configurations for serial-chain manipulators [23]. 2.3. Magnetorheological fluids Magnetorheological fluids (MRF) are so called “smart materials”. Applying an external magnetic field and directing it through the MRF changes the mechanical characteristics of the fluid within only a few milliseconds, from a low viscosity to an approximately solid state of the fluid. This behavior offers the possibility to lock as well as release the movement of the mechanic safety coupling. Since the amount of transmittable force through the MRF is dependent on the characteristics of the magnetic field, the fluid itself functions as an inherently safe coupling substance. Limiting the transmittable force to a safety threshold defined in ISO/TS 15066 by changing the magnetic field, enables a movement of the coupling before the collision is detected. MRF are used for example in technical applications such as mechanical dampers, clutches and breaks, utilizing their ability of a characteristic’s change within a very short period of time. [24, 25]Shafer and Kermani present in [26]and [27]the design of a two degrees of freedom safe manipulator with a single motor. The manipulator/robot uses magnetorheological fluids within its clutches to control the transmitted torque and therefore enable safe usage within human robot collaboration. Each joint of the robot combines two MRF-clutches, which shut the movement in opposite directions. The movement of each joint is controlled using the magnetic field and changing the behavior of the MRF within each clutch. The safety of the robot during HRC applications is ensured by limiting the transmittable torque by each of the robot’s joints.
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle/ Procedia CIRP 00 (2019) 000–000
910
3. Mechanic Safety Coupling Design Using the mechanical coupling unit reduces the transmitted kinetic energy from the robot system on the human worker in case of a collision, as already mentioned in Section 1. The human body is not capable of absorbing the high, transmitted kinetic energy, especially during a collision with a classical industrial robot, without potentially suffering serious injuries. Therefore, ISO/TS 15066 defines the acceptable limits for the transmitted energy based on the biomechanical threshold values for each region of the human body. Furthermore, the standard defines the calculation of the maximum relative velocity between the human worker and the robot. Since quasi-static contacts represent a significant higher hazard potential, designing a robot application with respect to quasi-static as well as transient collisions limits the maximum robot velocity and therefore the economic efficiency of the application. Following interrelationships are only valid for the transient contact. It is based on the assumption of the worstcase scenario, a complete inelastic contact situation, as shown in Figure 1. Here, the relative kinetic energy is being completely transferred to the affected body part.
Fig. 1. Contact model of a complete inelastic contact situation during a transient collision with mR being the effective mass of the robot depending on the robot position and movement, mH being the effective mass of the affected body part of the human worker, A being the area of contact between the robot and the part of the human body and vrel being the relative velocity between the robot and the affected body part.
The transmitted kinetic energy equals the change of the intrinsic energy as a result of the collision. Exceedance of the defined threshold values for the energy transmission is prohibited and must be taken into account during design of the HRC application. This value takes the maximum contact force Fmax as well as the stiffness k, depending on the affected part of the human body, into account: 𝐸𝐸 =
2 𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚
2𝑘𝑘
=
2 𝐴𝐴2 𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚
3
Specific values for mH for the application of equation (4) are listed in ISO/TS 15066.
Fig. 2. Simplified model of the mass distribution on a robot [17].
The effective mass of the robot mR depends on its actual position and movement, captured in the total mass M of all moved parts of the system, as well as the effective payload mL including the tool and the workpiece: 𝑚𝑚𝑅𝑅 =
𝑀𝑀 2
+ 𝑚𝑚𝐿𝐿
(5)
Heavy, classical industrial robots, in comparison to lightweight robots, are not slowed down significantly by the applied contact forces. Equations (2) to (4) show the possibility of increasing the maximum relative velocity of the robot during HRC applications by substantially reducing the effective mass of the robot, while respecting the maximum contact forces. Therefore, decoupling the heavy mass of the classical industrial robot from the effective mass of the tool as well as possible handling objects, during the braking process of the robot after a collision has been detected, shows a suitable approach for enabling usage of these robots during human robot collaboration. By decoupling the mass and enabling an almost complete free translational movement of the remaining collision mass (Fig. 3 a), only transient contact situations occur since a clamping of the human is not possible (Fig. 3 b).
(1)
2𝑘𝑘
Using the area of contact A between the robot and the body part, the maximum contact force Fmax can also be translated to a maximum contact pressure pmax. Comparing equation (1) with the complete kinetic energy at the center of gravity of the mass can be used to calculate the maximum acceptable relative velocity vrel: 𝐸𝐸 =
𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚 2 2𝑘𝑘
𝑣𝑣𝑟𝑟𝑟𝑟𝑟𝑟,𝑚𝑚𝑚𝑚𝑚𝑚 =
1
2 = 𝜇𝜇𝑚𝑚 𝑣𝑣𝑟𝑟𝑟𝑟𝑟𝑟,𝑚𝑚𝑚𝑚𝑚𝑚 2
𝐹𝐹𝑚𝑚𝑚𝑚𝑚𝑚
√𝜇𝜇𝑚𝑚𝑘𝑘
=
𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚 𝐴𝐴 √𝜇𝜇𝑚𝑚 𝑘𝑘
(2) (3)
Here µm captures the mass of the reduced two body system, consisting of the effective mass of the robot mR as well as the effective mass mH of the affected part of the human body, as also shown in Figure 2: 𝜇𝜇𝑚𝑚 = (
1
𝑚𝑚𝐻𝐻
+
1
𝑚𝑚𝑅𝑅
)
−1
(4)
Fig. 3. Mechanical safety coupling (a) with two translational degrees of freedom. Defined by the parallel kinematic of the coupling, plane E2 and plane E3 are always oriented parallel towards each other. The coupling can be mounted on the robot flange using the mechanical interface 2. The tool is connected using the mechanical interface 3.; (b) In case of a collision with the human worker (P), the robot’s (R) mass is decoupled from the tool and workpiece mass (W) which therefore prevents collisions with the full mass of the robotic system as well as quasi-static contacts at all.
4
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
For decoupling the mass of the robot from the mass of the tool within a very short time, specific actuators are needed. For this purpose, magnetorheological fluids (MRF) were identified as an inherent safe substance for actuating the mechanic coupling. The transmittable force and therefore the maximum load of the coupling is defined by the surface area covered with MRF. So far, the coupling is designed for a maximum payload of 10 kg in all translational degrees of freedom. Increasing the area covered with MRF also increases the maximum load of the coupling. The mechanic coupling itself is designed with three pairs of MRF actuators. These actuators function as ball joints and are connected with rigid bars (Fig. 4 a). This design also allows By varying the length of the bars, the braking distance of different robots can be compensated. Each ball joint is filled with magnetorheological fluid surrounding the ball itself to lock as well as release the movement of the joint. Since MRF are activated, i.e. change from a low viscosity fluid to an approximately solid state, by magnetic fields, the joint also contains several magnetic field cores. The cross section of a ball joint is given in Figure 4 b).
911
the magnetorheological fluid. Energizing each coil around a core separately and without a magnetic connection to other cores leads to misdirection and vortexes of the magnetic field. Since therefore a directed movement of the magnetic field lines through the MRF is not possible, a sufficient activation of the fluid doesn not take place. 4.2. Magnetic field bridges To overcome this problem of insufficient penetration of the MRF with the magnetic field, the magnetic cores next to each other are connected using a magnetic field bridge to create closed magnetic circuits (Fig. 6). Therefore, the complete magnetic flux is focused and directed through the MRF.
Fig. 5. Closed magnetic circuit within the ball joint, with the magnetic resistance of the MRF ℜ𝑀𝑀𝑀𝑀𝑀𝑀 , of the steel ball ℜ𝑠𝑠1 and the magnetic bridge ℜ𝑠𝑠2 , as well as the magnetic resistances of the cores ℜ𝑐𝑐1 , ℜ𝑐𝑐2 and the threaded rods ℜ𝑠𝑠𝑠𝑠1 and ℜ𝑠𝑠𝑠𝑠2 .
Fig. 4. Cross section of the ball joints of the mechanical safety coupling. (a) The coupling is designed using three pairs of ball joints each connected with a rigid bar (4). The ball itself (7) is surrounded by MRF (11). Balls in comparison to other geometries offer a bigger surface area (9) to be covered by the MRF which increases the maximum assignable force. Sealing rings (10) are used to prevent leakage of the fluid during regular usage. The magnetic field is generated with multiple magnetic field cores (12). (b) Enlarged display of the cross section of a ball joint.
The magnetorheological fluid itself should represent the highest magnetic resistance ℜ𝑀𝑀𝑀𝑀𝑀𝑀 within the circuit. As a result, the highest magnetic potential difference is present at the MRF. Simulative examination of the adjusted magnetic field design is shown in Figure 6.
4. Magnetic Field Design Besides the mechanical design of the coupling, the technical configuration of the magnetic fields is of superior importance for the functionality of the safety coupling. For this reason, simulation studies using ANSYS Workbench were carried out. The design of the coupling focusing on the magnetic design is described in the following. 4.1. Magnetic field cores and coils Overall, seven magnetic field cores are positioned around the ball to achieve a comprehensive penetration of the MRF with the magnetic field. Each magnetic field core is produced out of soft magnetic cobalt-iron alloy and wrapped with multiple layers of enamelled copper wire. The cores are mounted to the housing of the ball joint using threaded rods and are being oriented to realize a directed magnetic flux through
Fig. 6. Simulation of the magnetic flux density in the ball joint with adjusted magnetic field design. Since threaded rods are used to assemble the cores, the air within the threaded holes cut in the cores represents an area of high loss of the magnetic flux within simulation. This problem was fixed screwing the threaded rods completely into the core.
5. Experimental Studies and Validation To analyze the behavior of the coupling and the potential for integration in a robotic application, several collision experiments were performed. Goal of the experiments was to investigate the applied contact forces during a collision as well as the contact situation of the collision. The experimental setup and the results will be discussed in the following.
912
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle/ Procedia CIRP 00 (2019) 000–000
5
5.1. Experimental setup and collision procedure Two ball joints connected with a rod are used for the collision experiments. The upper ball joint is mounted on a UR5 lightweight robot. An acceleration sensor is used to detect a collision. All applied forces during the collision are recorded using a force measuring device especially designed for force examination within HRC applications. The complete experimental setup is shown in Figure 7. Beside the contact forces, the velocity of the robot during the collision experiments is recorded as well.
Fig. 7. Experimental setup for collision experiments using the mechanic safety coupling and a UR5 industrial robot by Universal Robots.
During the collision experiments, the robot first approaches a position in front of the force measuring device. Subsequently the robot accelerates until it has reached the previously defined velocity and collides with the force measuring device. In the process of a collision, the acceleration sensor sends a signal to the coupling control to deactivate the MRF as well as to the robot controller, which then stops the robot. The collision experiments are carried out under variation of the robot’s velocity. So far, the reaction time of the acceleration sensor and the time needed to decouple the mass of the robot from the mass of the tool by deactivation the MRF was not evaluated in the experiments, but will be part of future work.
Fig. 8. Force progression during collision for a robot velocity of vIR=900mm/s with multiple characteristic force impacts with decreasing intensity.
As already mentioned by Haddadin et. al. in [15], the applied forces during collision are highly dependent on the velocity of the robot. Therefore increasing the velocity of the robot leads to higher forces measured during the collision experiments. Figure 9 displays the maximum applied forces with standard deviation for different velocities of the robot. The increasing variance of the standard deviation is due to the manual positioning of the ball joint at the beginning of each collision experiment. Already minor torsion of the lower ball joint results in variable contact geometries and therefore in differences concerning the maximum applied forces. Overall the maximum measured forces during the collision experiments 𝐹𝐹𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 = 213 𝑁𝑁 did not exceed the lowest allowed force threshold of 𝐹𝐹𝑇𝑇,𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠ℎ = 220𝑁𝑁 (excluding the critical body parts face, forehead and head; here, collisions should be strictly avoided) defined by ISO/TS 15066, even for a robot velocity of 𝑣𝑣𝐼𝐼𝐼𝐼 = 900 𝑚𝑚𝑚𝑚⁄𝑠𝑠 .
5.2. Experimental results The force progressions – showed exemplary in Figure 8 – over all experiments indicate an exclusively transient contact situation under usage of the mechanical safety coupling. At the beginning of each force progression, the impact of the ball joint at the force measuring device is indicated by a single force pulse of a few milliseconds, followed by several short impacts with reduced force. This decreasing intensity of the applied force indicates a transmission of kinetic energy from the moved system on to the fixed force measuring device. Additional force impacts after the initial one are only observable from a robot velocity of 𝑣𝑣𝐼𝐼𝐼𝐼 = 250 𝑚𝑚𝑚𝑚⁄𝑠𝑠 upwards. This circumstance is explained by the mechanical design of the force measurement device. Below the mentioned velocity threshold, the force impact is too minor to compress the integrated spring enough to push away the ball joint and cause further impacts.
Fig. 9. Average maximum force with standard variation for all conducted collision experiments under variation of the robot velocity.
Using the measured maximum force during collision, an estimation of applicable robot velocities with integration of the mechanical safety coupling is possible. For this purpose, the measured forces can be matched with the maximum contact forces defined in ISO/TS 15066 for potentially affected body parts. The initial design of the coupling was developed to carry out first collision experiments. For this purpose, the coupling itself was positioned manually after each collision. Future designs will also have a mechanism to lock the coupling in an initial position to ensure accuracy.
6
Arik Lämmle et al. / Procedia CIRP 81 (2019) 908–913 Arik Laemmle / Procedia CIRP 00 (2019) 000–000
6. Discussion and Outlook In this paper, the development of a new mechanical safety coupling using magnetorheological fluids was evaluated. It was shown, that decoupling the effective mass of the robot from the effective mass of the tool as well as potentially handled parts during collision, significantly reduces the transmitted energy upon the affected parts of the human body. For this purpose, the MRF acts as an inherently safe coupling substance, since high external forces. e.g. as a result of a collision, already lead to a deformation of the fluid and therefore to a decoupling of the masses. The application of the mechanical safety coupling under real-world conditions was evaluated. The collision experiments point out an exclusively transient contact situation under usage of the mechanical safety coupling. This circumstance allows the design of the application with respect to the higher biomechanical thresholds for transient contact situations compared to quasi-static contacts defined in the standards. Furthermore, the robot is stopped solely by its own brakes and because of the decoupling of the robot mass and the mass of the tool not additionally through transmitting kinetic energy upon the collision partner. Using the safety coupling, the measured contact forces during the collision experiments did not exceed the defined force thresholds even for robot velocities beyond typical velocities of 500 mm/s in HRCapplications. Optimizing the magnetic field design, especially directing the movement of the magnetic field lines through the MRF and increasing the magnetic flux density within the MRF will be the focus of the author’s ongoing work. Furthermore, advanced collision experiments using classical industrial robots will be performed in the future. For this purpose a mechanically and magnetically optimized safety coupling, with increased payload will be developed. Since magnetorheological fluids tend to separate their individual ingredients, these experiments will also cover a detailed analysis of the long-term behavior of the mechanic safety coupling. References [1] Michalos, G., Makris, S., Tsarouchi, P., Guasch, T. et al., 2015. Design Considerations for Safe Human-robot Collaborative Workplaces 37, p. 248. [2] Santis, A. de, Siciliano, B., Luca, A. de, Bicchi, A. An atlas of physical human–robot interaction, in Mechanism and Machine Theory vol. 43, p. 253. [3] Lenz, C., Sotzek, A., Roder, T., Radrich, H. et al., 2011. Human workflow analysis using 3D occupancy grid hand tracking in a human-robot collaboration scenario, in 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, IEEE, p. 3375.
913
[4] Basso, F., Munaro, M., Michieletto, S., Pagello, E. et al., 2013. Fast and Robust Multi-people Tracking from RGB-D Data for a Mobile Robot, in Intelligent Autonomous Systems 12, Springer Berlin Heidelberg, Berlin, Heidelberg, p. 265. [5] Taipalus, T., Ahtiainen, J., 2011. Human detection and tracking with kneehigh mobile 2D LIDAR, in 2011 IEEE International Conference on Robotics and Biomimetics, IEEE, p. 1672. [6] Rio, M., Colas, F., Andries, M., Charpillet, F., 2016. Probabilistic sensor data processing for robot localization on load-sensing floors, in 2016 IEEE International Conference on Robotics and Automation (ICRA), IEEE, p. 4544. [7] Vogel, C., Fritzsche, M., Elkmann, N., 2016. Safe human-robot cooperation with high-payload robots in industrial applications, in 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI), IEEE, p. 529. [8] Kim, J., Alspach, A., Yamane, K., 2015. 3D printed soft skin for safe human-robot interaction, in 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, p. 2419. [9] Blue danube robotics. Technology for robots. https://www.bluedanuberobotics.com/airskin/. [10] Mittendorfer, P., Cheng, G. Uniform Cellular Design of Artificial Robotic Skin, in 2012 7th German Conference on Robotics, p. 1. [11] Mittendorfer, P., Cheng, G. Humanoid Multimodal Tactile-Sensing Modules, in IEEE Transactions on Robotics, vol. 27, no 3, 2011, p. 401. [12] Cirillo, A., Ficuciello, F., Natale, C., Pirozzi, S. et al. A Conformable Force/Tactile Skin for Physical Human–Robot Interaction, in 2016 IEEE Robotics and Automation Letters, vol. 1, no.1, p. 41. [13] Oberer-Treitz, S., Dietz, T., Verl, A., 2013. Safety in industrial applications: From fixed fences to direct interaction, in IEEE ISR 2013, IEEE, p. 1. [14] Haddadin, S., Albu-Schaffer, A., Frommberger, M., Hirzinger, G., 2008. The role of the robot mass and velocity in physical human-robot interaction - Part II: Constrained blunt impacts, in 2008 IEEE International Conference on Robotics and Automation, IEEE, p. 1339. [15] Haddadin, S., Albu-Schaffer, A., Frommberger, M., Rossmann, J. et al., 2009. The “DLR Crash Report”: Towards a standard crash-testing protocol for robot safety - Part I: Results, in 2009 IEEE International Conference on Robotics and Automation, IEEE, p. 272. [16] Deutsches Institut für Normung. DIN ISO/TS 15066 Roboter und Robotikgeräte - Kollaborierende Roboter. [17] Behrens, R., Elkmann, N., 2014. Study on meaningful and verified Thresholds for minimizing the consequences of human-robot collisions, in 2014 IEEE International Conference on Robotics and Automation (ICRA), IEEE, p. 3378. [18] Spaggiari, A. Properties and applications of Magnetorheological fluids, in 2012 Frattura ed Integrità Strutturale, 23, p. 48. [19] Wang, J., Meng, G. Magnetorheological fluid devices: Principles, characteristics and applications in mechanical engineering, in Proceedings of the IMechE, p. 165. [20] Yadmellat, P., Shafer, A.S., Kermani, M.R. Design and Development of a Single-Motor, Two-DOF, Safe Manipulator, in IEEE/ASME Transactions on Mechatronics, p. 1384. [21] Shafer, A.S., Kermani, M.R., 2009. On the feasibility and suitability of MR and ER based actuators in human friendly manipulators, in 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, IEEE, p. 2904.