Automatic Path Tracking and Target Manipulation of a ... - MDPI

micromachines Article

Automatic Path Tracking and Target Manipulation of a Magnetic Microrobot Jingyi Wang 1,2,4 , Niandong Jiao 1, *, Steve Tung 1,3 and Lianqing Liu 1, * 1 2 3 4

*

State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China; [email protected] University of Chinese Academy of Sciences, Beijing 100049, China Department of Mechanical Engineering, University of Arkansas, Fayetteville, AR 72701, USA; [email protected] Beijing Advanced Innovation Center for Imaging Technology, Capital Normal University, Beijing 100037, China Correspondence: [email protected] (N.J.); [email protected] (L.L.); Tel.: +86-24-2397-0540 (N.J.); +86-24-2397-0181 (L.L.)

Academic Editors: Duc Truong Pham and Nam-Trung Nguyen Received: 25 July 2016; Accepted: 18 November 2016; Published: 23 November 2016

Abstract: Recently, wireless controlled microrobots have been studied because of their great development prospects in the biomedical field. Electromagnetic microrobots have the advantages of control agility and good precision, and thus, have received much attention. Most of the control methods for controlling a magnetic microrobot use manual operation. Compared to the manual method, the automatic method will increase the accuracy and stability of locomotion and manipulation of microrobots. In this paper, we propose an electromagnetic manipulation system for automatically controlling the locomotion and manipulation of microrobots. The microrobot can be automatically controlled to track various paths by using visual feedback with an expert control algorithm. A positioning accuracy test determined that the position error ranges from 92 to 293 µm, which is less than the body size (600 µm) of the microrobot. The velocity of the microrobot is nearly proportional to the applied current in the coils, and can reach 5 mm/s. As a micromanipulation tool, the microrobot is used to manipulate microspheres and microgears with the automatic control method. The results verify that the microrobot can drag, place, and drive the microstructures automatically with high precision. The microrobot is expected to be a delicate micromachine that could play its role in microfluidics and blood vessels, where conventional instruments are hard to reach. Keywords: magnetic microrobot; automatic path tracking; expert control; magnetic microrobot; micromanipulation

1. Introduction Because of their small size, high controllability, and mobility, wireless controlled microrobots have promising development prospects in the field of survey and manipulation for a microscale space such as microfluidic systems, blood vessels, and inner tissues [1]. To ensure that microrobots have a locomotion ability, various actuation methods have been developed such as electromagnetic [2–7], thermal [8], chemical bubble [9], swimming tail [10,11], bacterial [12,13], and hybrid [14]. Among these actuation methods, the electromagnetic actuation method has characteristic advantages. These include high controllability, high accuracy, and a large actuating force as a result of changing the currents in the electromagnetic coils. In addition, the low electromagnetic field intensity used to control the microrobots will not cause harm to, or have any negative side effects on, human beings. However, the magnetic microrobots are mostly manually controlled through joystick or keyboard. The manual control method is an open-loop teleoperation in which the operator transmits motion Micromachines 2016, 7, 212; doi:10.3390/mi7110212

www.mdpi.com/journal/micromachines

Micromachines 2016, 7, 212

2 of 14

information to the microrobots through an input device, without any feedback [15]. With this method, the microrobot can be controlled to move in a microfluidic chip [16], in 3D space [17], as well as in vivo [18]. However, manually controlled microrobots can only correct off-track motion if instructed by their operator, and the effectiveness of operation is highly dependent on the experience of the operator. Moreover, the automatic control method for the microrobot cannot completely replace the manual control method. We believe that the automatic and the manual method are two modes of operation and can be applied in different situations respectively. Additionally, the automatic and manual method can compensate for each other. For example, the manual control method is not appropriate for tasks that require high repetition with high precision [19–21]. To overcome the shortcomings of the manual control method for microrobots, automatic control is necessary and indispensable. The automatic method is a path-tracking close-loop control method with visual feedback to transmit the positional information to the close-loop controller; by regulating the electromagnetic force in real time, the error between the microrobot and its target position can be eliminated. In the case of manipulating microrobots in vivo, the microrobots will have difficulties with the presence of perturbations, boundary effects, and so on. For example, in the condition of existing external disturbance such as uneven liquid surface or flowing liquid, the automatic control method can operate the microrobot with greater accuracy. Meanwhile, it is very difficult to operate the microrobot to do the repetitive motion with the manual method. Hence, with the automatic control method, the motion of the microrobot will be more robust with high repetition and high precision. In this paper, we propose the expert control algorithm for microrobot control with a novel electromagnetic manipulation system. Compared with proportional–integral–derivative (PID) controller [22,23], the expert controller has an advantage of self-adaption. Without testing the parameters of P, I and D, the expert controller can control microrobots with different sizes or magnetization. Also, with the expert control method, the driving force is reducing when the microrobot is nearing the target point. The tracking path is separated to several segments with target points. The moving speed is controlled to reduce to zero when the microrobot arrives at every target point. With the automatic control method using an expert control algorithm, the microrobot performs a point-by-point path-tracking motion. The predefined trajectory is divided into a series of adjacent points, then the microrobot tracks the points one by one to complete the path tracking. The microrobot can be controlled automatically to propel in the desired direction or track complex paths through image processing and a visual feedback control system. Furthermore, as a micromanipulation tool, the microrobot is used to drag, place, and rotate the micromachines automatically with high precision. 2. Methods 2.1. Theoretical Background of Magnetic Microrobot For the alignment of the microrobot on the xy plane, the generated magnetic field (B) and the following torque (T) is generated as: " T = VM × B = V

Mx My

#

"

×

Bx By

#

"

=V

Mx My

#

"

×

Bcosθ Bsinθ

# (1)

where V, M, and B denote the volume and magnetization vector of the microrobot and the magnetic flux vector of the external magnetic field on the xy plane, respectively. The alignment direction (θ) of B the microrobot is determined as tan θ = tan Byx . When manipulating the microrobot in fluidic environment, the viscous drag force is the major resistance force for transport. In order to manipulate the microrobot, the input magnetic force must overcome the resistance: F ≥ Fd (2)


3 of 14

where F is the magnetic force exerted on the microrobot, and Fd is the resistance of the fluid. The magnetic force is given by [24]: F = V (M · ∇)B (3) where V is the volume of the microrobot, M is the magnetization vector of the microrobot, and B is the magnetic flux vector of the external magnetic field. When the microrobot is actuated on the xy plane, ∇ is the gradient operator: "

∂ ∂x ∂ ∂y

∇=

# (4)

and the uniform magnetization of the microrobot and the magnetic flux are formed: " M=

Mx My

#

" ,B =

Bx By

# (5)

The magnetic force can be expressed in the following vector form: " F=

Fx Fy

#

"

=V

∂By ∂x ∂By ∂y

∂Bx ∂x ∂Bx ∂y ∂B

#"

Mx My

#

"

≈V

∂Bx ∂x

0

0

∂By ∂y

#"

Mx My

#



=V

∂B

x Mx ∂B ∂x



∂By ∂y



My

(6)

∂B

y y y ∂Bx ∂Bx x Compared with ∂B ∂x and ∂y , ∂x and ∂y have very small values: ~3% of ∂x and ∂y [25]. Mx and My are in same order of magnitude, hence the cross-term of the force matrix may be ignored and are set to zero. The magnetic field generated by the current flows through a spiral electromagnetic coil, and can be calculated by the Biot–Savart law as follows [26]:

B=

µ0 I 4π

dl × h

Z L

(7)

|h|3

where I is the current through the coil, L is the integral path, h is the full displacement vector from the wire element to the point where the field needs to be calculated, dl is a vector of the differential element of the current through the wire, and µ0 is the magnetic permeability (µ0 = 4π × 10−7 H/m). The magnetic fields generated by the coils at the arbitrary position (x, y) can be described as the multiplication of the electromagnetic field value per unit current Bˆ ( x, y) and the current (I) applied through each coil: B( x, y) = Bˆ ( x, y)I (8) The magnetic flux can be expressed as the function of the current I:  "

Bx By

#

"

=

Bˆ x,1 Bˆ y,1

Bˆ x,2 Bˆ y,2

Bˆ x,3 Bˆ y,3

Bˆ x,4 Bˆ y,4

#

   

I1 I2 I3 I4

    

(9)

where Bx ( In ) and By ( In ) (n = 1, 2, 3, 4) denote the x-axis and y-axis directional magnetic fields generated by the n-th coils. The magnetic force can be expressed as:


4 of 14

where B x ( I n ) and By ( I n ) (n = 1, 2, 3, 4) denote the x-axis and y-axis directional magnetic fields Micromachines 2016, 7, 212

4 of 14

generated by the n -th coils. The magnetic force can be expressed as: ∂B y   ∂B x    ∂B x  " # x ∂x  M Fx  # ∂B ∂B y M x  ∂B x x ∂ ∂ x  Mx ∂xx FF=x   = V  ∂x ∂ B∂x   =V y  M  y  Mx  M =∂BVy   ∂B x  F y=   V ∂By x  ∂∂B Fy My  y ∂y  My ∂By  y∂y ∂ y∂y  ∂y  ˆ ˆ ˆ ˆ I     ∂ Bˆ x ,1 ∂ Bˆ x , 2 ∂B ∂ˆB x , 4  1  x ,3 ˆ ∂ Bx,1 M x ∂ Bx,2 ∂ Bx,3 M M x ∂ Bx,4    M x Mx ∂x Mxx ∂x M x ∂x ∂x ∂x ∂ x   I 2  Mx ∂∂xx V =   V  ˆ ˆ ˆ ˆ B B B ∂∂BˆBˆyy,1 ∂ ∂ ∂ ˆ ˆ ∂B y , 4   I 3  ∂B y , 2 ∂ B y ,3 ,1 M yy y,4    M  MMy y ∂y Myy ∂yy,2 M Myy ∂yy,3 M ∂y∂ y ∂y ∂y ∂y I 4  

" F=

=

  I1 I2 I3 I4



(10) (10)

   

From Equation (10), the magnetic force, the magnetic flux, and the gradient magnetic fields used to propel the microrobot in any position, can be calculated using the current in every coil. Therefore, Therefore, the microrobot can be controlled to move to the desired position by changing the current in the electromagnetic coils. 2.2. Close-Loop Control of the Microrobot actuated automatically by using visual feedback control. In order actuate The microrobot microrobotcan canbebe actuated automatically by using visual feedback control. Intoorder to the microrobot to propeltoalong a desired accurately, a close-loop system using image processing actuate the microrobot propel along apath desired path accurately, a close-loop system using image and the expert algorithm is algorithm set up. Theisstructure diagram of the visual feedback expert control processing andcontrol the expert control set up. The structure diagram of the visual feedback is shown in Figure 1. in Figure 1. expert control is shown

Figure 1. Structure Structure diagram diagram of of the the visual visual feedback feedback control control of of the the microrobot. microrobot. The The error error between between the the Figure 1. real-time and target positions of the microrobot is set to be the input variable. B and B are the ∇ real-time and target positions of the microrobot is set to be the input variable. B and ∇B are the generated electromagnetic field, field, and and the the gradient gradient field field to to control control the the microrobot, microrobot,respectively. respectively. generated electromagnetic

First, the location information of the microrobot is identified by image recognition. First, the location information of the microrobot is identified by image recognition. Second, the target position is determined. Then, the distance and desired alignment angle Second, the target position is determined. Then, the distance and desired alignment angle between between the microrobot and the target position are calculated, where these are set to be the input the microrobot and the target position are calculated, where these are set to be the input variables. variables. The expert controller can be expressed as The expert controller can be expressed as U = f(E, I, K) U = f (E, I, K) where U is the output variable, E is the input set (distance and alignment angle between the microrobot and the target position), is theset output of the inference, K angle the knowledge item, and f is where U is the output variable, E is theI input (distance and alignment between the microrobot the intelligent operator of Expert Control. The principle of f is and the target position), I is the output of the inference, K the knowledge item, and f is the intelligent operator of Expert Control. The principle of f Kisthen If E and

where I is the result of the inference If of Einput E and K. Then, the required control behavior and K thenknowledge is exported according to I. The method of knowledge representation is genetic rule. The output of the inference generated the andknowledge knowledgeK.and experience (K). Then, the output where I is (I) theisresult of thethrough inference ofinput input(E) E and Then, the required control behavior (U) is generated according to the output of the inference. is exported according to I. The method of knowledge representation is genetic rule. The output of the Third, intelligentthrough operator f is the relational operator between the input the inference (I)the is generated the input (E) and knowledge and experience (K).variables Then, theand output gradient magnetic fields generated by of thethe unit current of each coil. The magnetic fields can be (U) is generated according to the output inference. calculated Equation (9). Third,from the intelligent operator f is the relational operator between the input variables and the The method of fields knowledge representation is genetic gradient magnetic generated by the unit currentrule. of each coil. The magnetic fields can be calculated from Equation (9).


5 of 14


5 of 14

The method of knowledge representation is genetic rule. The output of the inference (I) is generated through the input (E) and knowledge and experience (K). Then, the output (U) is generated according to to the the output output of of the the inference. inference. example,the thedistance distancebetween betweenmicrorobot microrobot and target point is calculated as input (E). For example, and thethe target point is calculated as input (E). The The output (U) is the current for the electromagnetic coils. The inference (I) is set as: output (U) is the current for the electromagnetic coils. The inference (I) is set as: If the distance (E) is zero, then set the output as as zero. zero. If the distance (E) is small (but not 0), then set the output output as as small. small. If the distance (E) is large, then set the output as large. large. The expression of zero, small, and large large of of distance distance can can be be expressed expressed as as EZ, EZ, ES, ES, and and EL. EL. Based on the knowledge and experience (K), EZ, ES, and EL can be expressed as: expressed as: EZ: E < 5 pixels EZ: E < 5 pixels ES: 5 pixels ≤ E ≤ 500 pixels ES: 5 pixels ≤ E ≤ 500 pixels EL: E > 500 pixels EL: E > 500 pixels Then, inference can be expressed as: Then, inference can be expressed as: If EZ, then U = 0;

ES,then thenUU= =0;f(E); If If EZ, EL, then MAX; If If ES, then UU = f=(E); f(E) can be set as a liner function of E. Or f(E)then can U be=set as a piecewise function. It can be adjusted If EL, MAX; according to the control precision. MAX can be set as a constant to control the microrobot at express f (E) can be set as a liner function of E. Or f (E) can be set as a piecewise function. It can be adjusted speed. according control precision. MAX can be set a constant topath-tracking control the microrobot at Using to thethe expert control algorithm, microrobot is aaspoint-by-point motion. With express speed. the expert control method, the driving force is reducing when the microrobot is nearing the target expertpath control algorithm,tomicrorobot is a point-by-point motion. With point.Using The the tracking is separated several segments with targetpath-tracking points. The moving speedthe is expert control method, the driving force is reducing when the microrobot is nearing the target point. controlled to reduce to zero when the microrobot arrives at every target point. This control method The is separated to severalofsegments withprocess target points. moving stability, speed is controlled can tracking meet thepath moderate requirements the control for theThe quickness, and nonto reduce to zero when the microrobot arrives at every target point. This control method can meet the overshoot. moderate requirements of the control process foralgorithm, the quickness, stability,the and non-overshoot. The function of knowledge, K, used in our is to change magnitude and direction The function of knowledge, K, used in our algorithm, is to change the magnitude and positions direction of the electromagnetic force, according to the distance between the real-time and the target of the force, according distance between theare real-time and the target positions of the electromagnetic microrobot. As shown in Figure to 2, the when the target points set as point A and point B, the of the microrobot. As shown in Figure 2, when the target points are set as point A and point B, the distance between the microrobot and point A is calculated first. Then, an electromagnetic force is distance between pointtoApoint is calculated first. decreases Then, an electromagnetic force is generated to drivethe themicrorobot microrobotand to move A. This force as the distance reduces, generated to drive thefalls microrobot to move point A. This forceatdecreases the distance reduces, and and the driving force to zero when thetomicrorobot arrives point A. as Then, the distance between the driving force falls to zero when the microrobot arrives at point A. Then, the distance between point point A and point B is recalculated and the electromagnetic force is regenerated. This algorithm A and point B is recalculated andaccurately the electromagnetic forceprecision, is regenerated. This algorithm and makes the makes the microrobot move with higher smaller overshoot, more microrobot move accurately with higher precision, smaller overshoot, and more robustness. robustness.

Figure 2. Schematic of expert control algorithm used in microrobot control. Figure 2. Schematic of expert control algorithm used in microrobot control.

Micromachines 2016, 7, 212 Micromachines 2016, 7, 212

6 of 14 6 of 14

3. 3. Experiments Experiments 3.1. Theoretical Background Magnetic Microrobot 3.1. Theoretical Background of of Magnetic Microrobot The electromagnetic manipulation microrobot was fabricated with NdFeB magnetic powder and The electromagnetic manipulation microrobot was fabricated with NdFeB magnetic powder and polydimethylsiloxane(PDMS). (PDMS). The NdFeB a high magnetization value,value, whichwhich can increase polydimethylsiloxane NdFeBpowder powderhas has a high magnetization can the propulsion force offorce the microrobot. increase the propulsion of the microrobot. The brief fabrication procedure microrobot are shown in Figure 3. First, the mold The brief fabrication procedure andand thethe microrobot are shown in Figure 3. First, the mold for thefor the microrobot was designed as a thin disk. The diameter and the thickness of the microrobot are microrobot was designed as a thin disk. The diameter and the thickness of the microrobot are 600 and 600μm, andrespectively. 200 µm, respectively. the was moldfabricated was fabricated an engraving machine (Benchtop 200 Second, Second, the mold usingusing an engraving machine (Benchtop Engravers EGX-600/400, Roland DGA Corporation, Irvine, CA, USA) equipped with a milling cutter Engravers EGX-600/400, Roland DGA Corporation, Irvine, CA, USA) equipped with a milling cutter with a diameter of 200 µm. The mold was carved on an acrylic plate. Third, the PDMS and NdFeB with a diameter of 200 μm. The mold was carved on an acrylic plate. Third, the PDMS and NdFeB magnetic powder were mixed a weight ratio 1:1. order make the PDMS and curing agent magnetic powder were mixed at at a weight ratio ofof 1:1. InIn order toto make the PDMS and curing agent sufficientand andeven-mixing, even-mixing,the thePDMS PDMSand and curing curing agent agent were mixed sufficient mixed first firstfor forabout about55min, min,and andthen thenthe magnetic powder was added. Finally, the mixture was stirred forfor 3030 min. Fourth, the mixture was put the magnetic powder was added. Finally, the mixture was stirred min. Fourth, the mixture was into thethe mold andand the redundant mixture was removed. The microrobot was degassed in a vacuum put into mold the redundant mixture was removed. The microrobot was degassed in a ◦ C for 4 h. Finally, the microrobot box to remove bubbles, and then placed in the oven to bake at 60 vacuum box to remove bubbles, and then placed in the oven to bake at 60 °C for 4 h. Finally, the was removed the mold tweezers it wasafter cured. microrobot was from removed fromwith the mold withafter tweezers it was cured.

Figure 3. 3.Process of of microrobot fabrication. (a) (a) Curving thethe microrobot mold onon anan acrylic plate; Figure Process microrobot fabrication. Curving microrobot mold acrylic plate; (b)(b) putting the microrobot into the vacuum chamber to remove bubbles; (c) baking the microrobot forfor putting the microrobot into the vacuum chamber to remove bubbles; (c) baking the microrobot 4 h4 at 60 °C; (d) removing the microrobot from the mold; (e) top view of the microrobot. ◦ h at 60 C; (d) removing the microrobot from the mold; (e) top view of the microrobot.

3.2. Design of the Electromagnetic Manipulation System 3.2. Design of the Electromagnetic Manipulation System As shown in Figure 4, there are six electromagnetic coils in the electromagnetic manipulation As shown in Figure 4, there are six electromagnetic coils in the electromagnetic manipulation system to manipulate the microrobot in three dimensions. Two of the pairs of the electromagnetic system to manipulate the microrobot in three dimensions. Two of the pairs of the electromagnetic coils were set to be orthogonal in the horizontal plane, whereas the other pair was set in the vertical coils were set to be orthogonal in the horizontal plane, whereas the other pair was set in the vertical plane. The horizontal electromagnetic coils were installed on the height-adjustable pillar. Every plane. The horizontal electromagnetic coils were installed on the height-adjustable pillar. Every electromagnetic coil is supported by the nylon plastic frames, which were designed for forward and electromagnetic coil is supported by the nylon plastic frames, which were designed for forward and backward precession. The height-adjustable pillar and the nylon plastic frames are convenient to backward precession. The height-adjustable pillar and the nylon plastic frames are convenient to adjust the vertical and horizontal position of the electromagnetic coils, respectively. The manipulation adjust the vertical and horizontal position of the electromagnetic coils, respectively. The manipulation region of the system is limited as a result of the constant structure. To overcome this disadvantage, region of the system is limited as a result of the constant structure. To overcome this disadvantage, we we propose a novel electromagnetic manipulation system that has position-adaptable propose a novel electromagnetic manipulation system that has position-adaptable electromagnetic electromagnetic coils. By regulating the six-position adaptable electromagnetic coils, the coils. By regulating the six-position adaptable electromagnetic coils, the manipulation area can be manipulation area can be adjusted to expand or shrink. A long-focus microscope with a chargeadjusted to expand or shrink. A long-focus microscope with a charge-coupled device (CCD) was used coupled device (CCD) was used to observe and provide visual feedback through the upper to observe and provide visual feedback through the upper electromagnetic coil in the z direction. electromagnetic coil in the z direction.


7 of 14


7 of 14


7 of 14

Figure diagram of the manipulation system.system. There areThere six positionFigure 4. 4. Schematic Schematic diagram of electromagnetic the electromagnetic manipulation are six Figure 4. Schematic diagram of the electromagnetic manipulation system. There are six positionadjustable electromagnetic coils to adjust the manipulation region. Each electromagnetic coilcoil is position-adjustable electromagnetic coils to adjust the manipulation region. Each electromagnetic adjustable electromagnetic coilsframes, to adjust thewere manipulation region. Each electromagnetic coil is supported by the nylon plastic which designed as forward and backward precession. is supported by the nylon plastic frames, which were designed as forward and backward precession. supported by the nylon plastic frames, which wereframes designed as forward and backward precession. The height-adjustable pillar nylon plastic convenient to adjust the vertical vertical and The height-adjustable pillar and and the the nylon plastic frames are are convenient to adjust the and The height-adjustable pillar and the nylon plastic frames are convenient to adjust the vertical and horizontal position of the electromagnetic coils, respectively. horizontal position of the electromagnetic coils, respectively. horizontal position of the electromagnetic coils, respectively.

The actual electromagnetic manipulation system used is shown in Figure 5. It consists of an array The actual actualelectromagnetic electromagneticmanipulation manipulationsystem system used is shown in Figure It consists of an The used is shown in (Navitar Figure 5. 1-62317, It 5. consists of an array of iron-core electromagnetic coils, a motorized long-focus microscope Navitar Inc., array of iron-core electromagnetic coils, a motorized long-focus microscope (Navitar 1-62317, Navitar of iron-core electromagnetic coils,acquisition a motorized long-focus microscope 1-62317, Navitar Inc., New York, NY, USA), two data cards (NI PCI-6229 DAQ(Navitar Card, National Instruments Inc., New NY, USA), two data acquisition cards (NIPCI-6229 PCI-6229DAQ DAQCard, Card, National Instruments Instruments New York,York, NY, USA), two data cards (NI Corporation, Austin, TX, USA), aacquisition direct current (DC) power supply, a micron National 3D positioning stage, Corporation, Austin, Austin, TX, TX, USA), USA), aa direct direct current current (DC) (DC) power power supply, supply, a micron micron 3D 3D positioning positioning stage, stage, Corporation, and a power amplifier. The 3D positioning stage is used to adjust thea position of the microrobot in and a The 3D 3D positioning positioning stage stage is is used used to to adjust adjust the the position position of of the the microrobot and a power power amplifier. amplifier. The microrobot in in the container. the container. the container. The electromagnetic coils are supported by nylon plastic frames, which will not affect the The electromagnetic coils areare supported by nylon plastic frames, which will notwill affectnot the magnetic The electromagnetic supported byconductive nylon plastic frames, which the magnetic field as a result coils of their nonmagnetic property. The magnetic fieldaffect for unit field as a result of their nonmagnetic conductive property. The magnetic field for unit current of the magnetic as a result of their conductive property. The magnetic fieldcurrent for unit current of field the electromagnetic coils nonmagnetic is 2871 A/m, and the magnetic gradient field for unit is 2. electromagnetic coils is 2871 A/m, and the A/m, magnetic gradient fieldgradient for unit field current isunit 95726 A/mis 2 current of the electromagnetic coils is 2871 and the magnetic for current 95726 A/m . The detailed specification of the electromagnetic coils is described in Table 1. The detailed specification of the electromagnetic coils is described 1. in Table 1. 95726 A/m2. The detailed specification of the electromagnetic coils in is Table described

Figure 5. The whole electromagnetic manipulation system. Figure 5. The whole electromagnetic manipulation system. Figure 5. The whole electromagnetic manipulation system.


8 of 14


8 of 14

Table 1. 1. Parameters Parameters of of electromagnetic electromagnetic coils. coils. Table

Coils Coils Horizontal Horizontal Vertical Vertical

Coils Insider Radius/Outsider Insider Radius/Outsider Coils Turns Radius (mm) Turns Radius (mm) 500 500 55/60 55/60 55/60 500 500 55/60

Diameter ofCopper Copper Diameter of Wires Wires(mm) (mm) 0.51 0.51 0.51 0.51

Resistance Iron Resistance Iron Cores (Ω)(Ω) Cores 16 16 16 16

Yes Yes No No

In In order order to to manipulate manipulate the the microrobot, microrobot, the the location location information information is is acquired acquired by by the the electrical electrical forcing microscope. The control algorithm, programmed by LabVIEW (Labview 2012, forcing microscope. The control algorithm, programmed by LabVIEW (Labview 2012, National National Instruments Corporation), processes this information, generates the required output signal, and Instruments Corporation), processes this information, generates the required output signal, and then then transmits transmits it it to to the the data data acquisition acquisition card. card. The The required required output output current current values, values, amplified amplified by by the the power power amplifier, applied to amplifier, are are applied to the the electromagnetic electromagnetic manipulating manipulating system system to to generate generate the the required required magnetic magnetic field. The microrobot is placed in a cubic container with deionized (DI) water on the micron field. The microrobot is placed in a cubic container with deionized (DI) water on the micron 3D 3D positioning positioning stage stage in in the the center center of of the the electromagnetic electromagnetic coils. coils. The The electromagnetic electromagnetic manipulation manipulation system system can be used used to toobserve observeand andaccurately accuratelymanipulate manipulate microrobot manually, as our previous can be thethe microrobot manually, as our previous workwork [16], [16], or automatically. or automatically. 4. Results Results Velocity Measurement 4.1. Velocity In order thethe velocity of the various magnetic gradient fields were applied. order totomeasure measure velocity of microrobot, the microrobot, various magnetic gradient fields were These fields canfields be changed by the applied coils. the coils. applied. These can be changed by thecurrent appliedthrough current the through The microrobot microrobot was was placed placed on on the the surface surface of of DI DI water. Initially, Initially, it started to accelerate from its The stationary state state due due to the magnetic force. With With the the rapid rapid increase increase in in microrobot microrobot speed, speed, the the fluid stationary viscous force was instantaneously increased. The The microrobot microrobot can can instantaneously instantaneously reach reach a uniform viscous velocity in a fluidic environment [27]. Five Five experiments experiments were conducted at each applied current, and As shown shown in in Figure Figure 6, 6, the the velocity velocity of of the the microrobot microrobot increased increased the average velocities were calculated. As with an increase in applied current in the coils. From the experimental experimental results, results, we we find find that that the the velocity velocity of the microrobot is approximately approximately proportional to to the applied applied current current in the coils. When When the the applied applied current current is is 11 A, A, the velocity velocity of of the proportional microrobot can be greater than 5 mm/s. mm/s. notnot permanent. TheThe microrobot is first accelerated and then The velocity velocity in inthe thetracking trackingprocess processis is permanent. microrobot is first accelerated and moves with awith constant velocity. Finally,Finally, it decelerates gradually. In order to then moves a constant velocity. it decelerates gradually. In track orderthe to path trackaccurately, the path we set this we initial driving current 0.5 A (this initial is set atisabout 2 mm/s). accurately, set this initial drivingatcurrent at 0.5means A (thisthe means thevelocity initial velocity set at about 2 mm/s).

Figure 6. Velocity Velocity measurement measurement ofof the the microrobot microrobot according according to to the the current current through through the electromagnetic coils.

4.2. Performance of the Close-Loop Control The position accuracy of the microrobot is an important characteristic for close-loop control. The microrobot is controlled to track four paths (A, B, C, D) to evaluate the control accuracy, as shown in


9 of 14

4.2. Performance of the Close-Loop Control The position accuracy of the microrobot is an important characteristic for close-loop control. Micromachines 2016, 7, 212 9 of 14 The microrobot is controlled to track four paths (A, B, C, D) to evaluate the control accuracy, as shown in Figure The permissibleerror erroraffects affectsthe theperformance performanceof ofthe thecontrol. control. Based Based on on aa number of previous Figure 7. 7. The permissible previous experiments, experiments, close-loop control of the microrobot is characterized by high accuracy, good stability, and and rapid rapid response response when when the the permissible permissible error error range range is is set set at at 10 10pixels. pixels. Therefore, Therefore, the the permissible permissible error 500 µm in in length. This is less than thethe diameter of error range rangeisisset setatat10 10pixels, pixels,which whichcorresponds correspondstoto 500 μm length. This is less than diameter the microrobot. of the microrobot.

Figure 7. 7. Position microrobot. (a) Tracking lineslines (14 mm) for position error Figure Positionerror errormeasurement measurementof of microrobot. (a) Tracking (14 mm) for position measurement; (b) position error measurement in pixels and average time versus different initial error measurement; (b) position error measurement in pixels and average time versus different velocities. initial velocities.

As shown in Figure 7a, four equilong straight paths (A, B, C, D, 14 mm) were established for the As shown in Figure 7a, four equilong straight paths (A, B, C, D, 14 mm) were established for the microrobot to track. The microrobot was propelled along these paths, back and forth, with an microrobot to track. The microrobot was propelled along these paths, back and forth, with an automatic automatic control method. Then, the position errors were calculated to measure the control control method. Then, the position errors were calculated to measure the control performance. With performance. With different initial velocities (5 mm/s, 4 mm/s, 3 mm/s, 2 mm/s), each path is tracked different initial velocities (5 mm/s, 4 mm/s, 3 mm/s, 2 mm/s), each path is tracked three times, and three times, and the errors and average time are shown in Figure 7b. The position errors of the the errors and average time are shown in Figure 7b. The position errors of the tracking path range tracking path range from 92 to 293 μm. Consequently, the errors are all within the preset threshold from 92 to 293 µm. Consequently, the errors are all within the preset threshold value, and smaller than value, and smaller than the half-size of the microrobot. As shown in Figure 7b, the position errors the half-size of the microrobot. As shown in Figure 7b, the position errors become bigger as the initial become bigger as the initial velocity increases. The time that the microrobot completes the straight velocity increases. The time that the microrobot completes the straight line increases when the initial line increases when the initial velocity increases. Hence, the microrobot has a good position accuracy velocity increases. Hence, the microrobot has a good position accuracy when the initial velocity is set when the initial velocity is set at a small value. In order to perform great accuracy, the initial velocity at a small value. In order to perform great accuracy, the initial velocity is set at small initial velocity is set at small initial velocity when the microrobot performs close-loop tracking. when the microrobot performs close-loop tracking. 4.3. Close-Loop Tracking Performance of Microrobot 4.3. Close-Loop Tracking Performance of Microrobot Based on the expert control algorithm using visual feedback, a number of experiments regarding Based on the expert control algorithm using visual feedback, a number of experiments regarding tracking on different paths were executed. First, the microrobot was recognized and the position tracking on different paths were executed. First, the microrobot was recognized and the position information was acquired. Then, different paths were drawn for the microrobot tracking: rectangular, information was acquired. Then, different paths were drawn for the microrobot tracking: rectangular, triangular, hexagonal, square spiral, and customized paths. When these are determined, the triangular, hexagonal, square spiral, and customized paths. When these are determined, the microrobot microrobot can accurately propel along the desired paths automatically. As shown in Figure 8, the can accurately propel along the desired paths automatically. As shown in Figure 8, the green paths green paths were designed as the desired paths for the microrobot. The red arrow indicates the were designed as the desired paths for the microrobot. The red arrow indicates the direction of direction of movement. The movements of the microrobot along the rectangular, triangular, movement. The movements of the microrobot along the rectangular, triangular, hexagonal, square hexagonal, square spiral, and customized paths takes approximately 35, 32, 40, 70 and 61 s, spiral, and customized paths takes approximately 35, 32, 40, 70 and 61 s, respectively. respectively.


10 of 14

Figure 8. The expert control tracking performance of the microrobot. The green paths are set for the microrobot to track. The movement directions of the microrobot are marked in red dotted lines. The microrobot tracks the rectangular path (a); triangular path (b); hexagonal path (c); square spiral path (d); and complicated customized path (e). The scale bar is 5 mm.

4.4. Manipulating Microspheres The microrobot can be controlled to manipulate zirconia microspheres on the liquid surface. The diameter of the microsphere is approximately 1 mm, and it is placed on the liquid surface at random. The microrobot is controlled to propel towards the microsphere, and when it makes contact, it attaches to the microsphere as a result of the tensile force. As shown in Figure 9, for easily observation two points (A and B) are marked on the substrate, and the distance between them is 10 mm. In Figure 9, the microrobot is controlled to move towards the microsphere (a). The microrobot drags the microsphere and moves to A point automatically (b). On arrival at A point (c), it drags the microsphere (d) and moves to B point automatically (e). Then, it is controlled to accelerate and separate from the microsphere (f). The green paths are set for the microrobot to track. The movement directions are marked by red dotted arrow lines. The procedure of manipulation with one microsphere takes approximately 25 s.


11 of 14

We can also control the microrobot to manipulate two microspheres. In Figure 10, two microspheres are placed on the liquid surface randomly (a). The microrobot is controlled to manipulate the first microsphere to the desired point A automatically (b). Then, it accelerates to leave the Micromachines 2016, 7, 212 11 of 14 first microsphere and moves to the second (c). The microrobot manipulates the second microsphere (d) and Micromachines 2016, 212 automatically moves to point B (e). Finally, it leaves the second microsphere 11 of 14 microsphere (d) 7,and automatically moves to point B (e). Finally, it leaves the second microsphere at point B (f) and at moves point B (f) and moves on. The two microspheres are patterned at point A and B successfully. The on. The two microspheres are patterned at point ABand B successfully. Thesecond manipulation procedure microsphere and automatically to point (e).approximately Finally, it leaves microsphere at manipulation(d) procedure with two moves microspheres takes 50 the s. The experiment results with two microspheres takes approximately 50 s. The experiment results show that the microrobot can point B (f) and moves on. The two microspheres are patterned at point A and B successfully. The show that the microrobot can drag, place, and pattern the microspheres automatically with high drag,manipulation place, and pattern the with microspheres automatically with high precision. procedure two microspheres takes approximately 50 s. The experiment results precision. show that the microrobot can drag, place, and pattern the microspheres automatically with high precision.

Figure 9. Manipulation of the microrobot with one microsphere. The microrobot and zirconia

Figure 9. Manipulation of the microrobot with one microsphere. The microrobot and zirconia microsphere are marked in red circles. The green paths are set for the microrobot to track. The microsphere marked in red circles. The green paths are set for the microrobot to track. Figure 9. are Manipulation microrobot one by microsphere. microrobot movement directions of of thethe microrobot are with marked red dotted The arrow lines. (a) and The zirconia zirconia The movement directions of the microrobot are marked by red dotted arrow lines. (a) The zirconia microsphere are marked in red circles. The green paths are set for the microrobot to track. The microspheres are placed on the liquid surface at random. (b) The microrobot drags the microsphere movement directions of the microrobot are marked by red dotted arrow lines. (a) The zirconia microspheres are placed on the liquid surface at random. (b) The microrobot drags the microsphere and moves to A point automatically. On arrival at A point (c), it drags the microsphere (d) and moves microspheres are placed on the liquid surface microrobot drags microsphere and moves to A point automatically. arrivalat atrandom. A accelerate point(b) (c),The it drags the from microsphere (d) and(f). moves to B point automatically (e). Then, itOn is controlled to and separate the the microsphere and moves to A point automatically. On arrival at A point (c), it drags the microsphere (d) and moves to B point (e). Then, it is controlled accelerate and separate microsphere (f). Then, automatically it leaves the zirconia microsphere at point B.toThe manipulation procedurefrom takesthe approximately point automatically (e). Then, it is controlled to accelerate and separate from the microsphere (f). 25itBs.leaves Then,to the zirconia microsphere at point B. The manipulation procedure takes approximately Then, it leaves the zirconia microsphere at point B. The manipulation procedure takes approximately 25 s. 25 s.

Figure 10. Manipulation of the microrobot with two microspheres. Two microspheres are placed on the liquid surface randomly (a). The microrobot is controlled to manipulate the first microsphere to Figure 10. Manipulation of the microrobot with two microspheres. Two microspheres are on the10. desired point A automatically (b). Then, it accelerates to leave the first andplaced moves to on Figure Manipulation of the microrobot with two microspheres. Twomicrosphere microspheres are placed the liquid surface randomly (a). The microrobot is controlled to manipulate the first microsphere the second (c). randomly The microrobot manipulates the second microsphere (d) and automatically moves to to to the liquid surface (a). The microrobot is controlled to manipulate the first microsphere the desired A automatically (b). Then, it accelerates to leave first microsphere and moves to point B (e). point Finally, it leaves the second microsphere at point B (f)the and moves on. The manipulation the desired point automatically (b). Then, itthe accelerates to leave the moves to the second (c).AThe microrobot manipulates second microsphere (d)first and microsphere automatically and moves to procedure takes approximately 50 s. the second microrobot manipulates the second microsphere andon. automatically moves to point B(c). (e). The Finally, it leaves the second microsphere at point B (f) and (d) moves The manipulation point B (e). Finally, it leaves the second procedure takes approximately 50 s. microsphere at point B (f) and moves on. The manipulation 4.5. Manipulating Microgear

procedure takes approximately 50 s. The microrobot can be controlled to manipulate a red PDMS gear with four teeth. The external 4.5. Manipulating Microgear diameter of the gear is 2.25 mm. When the microrobot contacts the gear, it attaches to the gear as a The microrobot can be controlled to manipulate a red PDMS gear with four teeth. The external diameter of the gear is 2.25 mm. When the microrobot contacts the gear, it attaches to the gear as a


12 of 14

4.5. Manipulating Microgear Micromachines 2016, 7, 212

12 of 14

The microrobot can be controlled to manipulate a red PDMS gear with four teeth. The external diameter of the gear is 2.25 mm. When microrobot contacts the it attaches to the as a result of the tensile force. In Figure 11, thethe microrobot is controlled to gear, do a circular motion bygear tracking result of the tensile force. Figurecircle). 11, theHence, microrobot is controlled to do circular motionOne by tracking a predefined circular pathIn(green the microrobot pulls thea gear to rotate. tooth of athe predefined circular path (green circle). Hence, the microrobot pulls the gear to rotate. One gear is stroked in yellow to observe the rotating process easily. The rotational speed of thetooth gearof is the gear is stroked in yellow to observe the rotating process easily. The rotational speed of the gear is approximately 6 rpm. The experiment result shows that the microrobot can manipulate approximately 6 rpm. The experiment result shows that microrobot can manipulate microstructures microstructures with high precision and flexibility. Thethe manual control method is difficult to achieve with high precision and flexibility. The manual control method is difficult to achieve rotating the gear. rotating the gear. The experiment result also proves the effectiveness of the proposed automatic The experiment result also proves the effectiveness of the proposed automatic control method. control method.

Figure 11. 11. Rotate Rotatea agear gear microrobot. microrobot the to gear to by rotate by tracking Figure byby thethe microrobot. TheThe microrobot pullspulls the gear rotate tracking a circulara circular path. The rotational speed of the gear is approximately 6 rpm. path. The rotational speed of the gear is approximately 6 rpm.

5. Conclusions 5. Conclusions This paper proposed an automatic control method for a microrobot by means of an This paper proposed an automatic control method for a microrobot by means of an electromagnetic electromagnetic manipulation system. The microrobot was designed as a thin disk with diameter 600 manipulation system. The microrobot was designed as a thin disk with diameter 600 µm, which is μm, which is fabricated from a mixture of PDMS and NdFeB magnetic powder. The electromagnetic fabricated from a mixture of PDMS and NdFeB magnetic powder. The electromagnetic manipulation manipulation system was designed with six position-adaptable electromagnetic coils, which are used system was designed with six position-adaptable electromagnetic coils, which are used to adjust to adjust the manipulation region. The microrobot can be controlled to move in the manipulation the manipulation region. The microrobot can be controlled to move in the manipulation region region automatically by using visual feedback with an expert control algorithm. The positioning automatically by using visual feedback with an expert control algorithm. The positioning accuracy accuracy is tested, and the position error ranges from 92 to 293 μm, which is less than half the body is tested, and the position error ranges from 92 to 293 µm, which is less than half the body size of size of the microrobot. The effectiveness of the microrobot automatic movement was verified by the microrobot. The effectiveness of the microrobot automatic movement was verified by driving the driving the microrobot to track various paths (rectangular, triangular, hexagonal, square spiral and microrobot to track various paths (rectangular, triangular, hexagonal, square spiral and a complicated a complicated customized path). As a micromanipulation tool, the microrobot is used to manipulate customized path). As a micromanipulation tool, the microrobot is used to manipulate microspheres microspheres with diameter 1 mm. The experimental results verify that the microrobot can drag, with diameter 1 mm. The experimental results verify that the microrobot can drag, place, and drive place, and drive the microstructures automatically with high precision. As a result of its small size the microstructures automatically with high precision. As a result of its small size and high-precision and high-precision automatic control, the microrobot is expected to be a delicate instrument that automatic control, the microrobot is expected to be a delicate instrument that could be used for could be used for microfluidic control, intravascular survey, drug delivery, and minimal invasive microfluidic control, intravascular survey, drug delivery, and minimal invasive treatment. treatment. Acknowledgments: This paper is supported by National Natural Science Foundation of China (Grant No. 61573339) and theThis CAS/SAFEA International Program for Creative Research Teams.(Grant No. Acknowledgments: paper is supported by Partnership National Natural Science Foundation of China 61573339) and the CAS/SAFEA International Partnership Program for Creative Research Teams. Author Contributions: J.W. and N.J. conceived and designed the experiments; J.W. fabricated the devices and performed the experiments; J.W. N.J. and conceived S.T. analyzed data; J.W. N.J. designed set up the the devices experiment Author Contributions: J.W. and andthe designed theand experiments; J.W. and fabricated and system; J.W. wrote the paper; L.L. supervised the entire work. performed the experiments; J.W. and S.T. analyzed the data; J.W. and N.J. designed and set up the experiment Conflicts of Interest: authors no conflict of interest. system; J.W. wrote theThe paper; L.L. declare supervised the entire work.

Conflicts of Interest: The authors declare no conflict of interest.

References 1.

Nelson, B.J.; Kaliakatsos, I.K.; Abbott, J.J. Microrobots for minimally invasive medicine. Annu. Rev. Biomed. Eng. 2010, 12, 55–85.


13 of 14

References 1. 2. 3.

4.

5.

6. 7. 8. 9. 10.

11. 12.

13. 14.

15.

16. 17. 18. 19. 20. 21.

Nelson, B.J.; Kaliakatsos, I.K.; Abbott, J.J. Microrobots for minimally invasive medicine. Annu. Rev. Biomed. Eng. 2010, 12, 55–85. [CrossRef] [PubMed] Go, G.; Choi, H.; Jeong, S.; Lee, C.; Ko, S.Y.; Park, J.-O.; Park, S. Electromagnetic navigation system using simple coil structure (4 Coils) for 3-D locomotive microrobot. IEEE Trans. Magn. 2015, 51, 1–7. Tottori, S.; Zhang, L.; Qiu, F.; Krawczyk, K.K.; Franco-Obregón, A.; Nelson, B.J. Magnetic helical micromachines: Fabrication, controlled swimming, and cargo transport. Adv. Mater. 2012, 24, 811–816. [CrossRef] [PubMed] Kim, S.; Qiu, F.; Kim, S.; Ghanbari, A.; Moon, C.; Zhang, L.; Nelson, B.J.; Choi, H. Fabrication and characterization of magnetic microrobots for three-dimensional cell culture and targeted transportation. Adv. Mater. 2013, 25, 5863–5868. [CrossRef] [PubMed] Zhang, Y.; Yue, M.; Guo, D.; Wang, D.; Yu, H.; Jiang, S.; Zhang, X. Characteristics of spatial magnetic torque of an intestine capsule micro robot with a variable diameter. Sci. China Ser. E Technol. Sci. 2009, 52, 2079–2086. [CrossRef] Diller, E.; Floyd, S.; Pawashe, C.; Sitti, M. Control of multiple heterogeneous magnetic microrobots in two dimensions on nonspecialized surfaces. IEEE Trans. Robot. 2012, 28, 172–182. [CrossRef] Chowdhury, S.; Jing, W.; Cappelleri, D.J. Controlling multiple microrobots: Recent progress and future challenges. J. Micro-Bio Robot. 2015, 10, 1–11. [CrossRef] Shen, X.; Viney, C.; Johnson, E.R.; Wang, C.; Lu, J.Q. Large negative thermal expansion of a polymer driven by a submolecular conformational change. Nat. Chem. 2013, 5, 1035–1041. [CrossRef] [PubMed] Búzás, A.; Kelemen, L.; Mathesz, A.; Oroszi, L.; Vizsnyiczai, G.; Vicsek, T.; Ormos, P. Light sailboats: Laser driven autonomous microrobots. Appl. Phys. Lett. 2012, 101, 041111. [CrossRef] Hwang, G.; Braive, R.; Couraud, L.; Cavanna, A.; Abdelkarim, O.; Robert-Philip, I.; Beveratos, A.; Sagnes, I.; Haliyo, S.; Régnier, S. Electro-osmotic propulsion of helical nanobelt swimmers. Int. J. Robot. Res. 2011, 30, 806–819. [CrossRef] Huang, C.; Lv, J.-A.; Tian, X.; Wang, Y.; Yu, Y.; Liu, J. Miniaturized swimming soft robot with complex movement actuated and controlled by remote light signals. Sci. Rep. 2015, 5, 17414. [CrossRef] [PubMed] Martel, S.; Felfoul, O.; Mathieu, J.-B.; Chanu, A.; Tamaz, S.; Mohammadi, M.; Mankiewicz, M.; Tabatabaei, N. Mri-based medical nanorobotic platform for the control of magnetic nanoparticles and flagellated bacteria for target interventions in human capillaries. Int. J. Robot. Res. 2009, 28, 1169–1182. [CrossRef] [PubMed] Martel, S.; Tremblay, C.C.; Ngakeng, S.; Langlois, G. Controlled manipulation and actuation of micro-objects with magnetotactic bacteria. Appl. Phys. Lett. 2006, 89, 233904. [CrossRef] Li, D.; Choi, H.; Cho, S.; Jeong, S.; Jin, Z.; Lee, C.; Ko, S.Y.; Park, J.O.; Park, S. A hybrid actuated microrobot using an electromagnetic field and flagellated bacteria for tumor-targeting therapy. Biotechnol. Bioeng. 2015, 112, 1623–1631. [CrossRef] [PubMed] Shull, P.; Niemeyer, G. Open-loop bilateral teleoperation for stable force tracking. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, St. Louis, MO, USA, 10–15 October 2009; pp. 5121–5126. Wang, J.; Jiao, N.; Tung, S.; Liu, L. Magnetic microrobot and its application in a microfluidic system. Robot. Biomim. 2014, 1, 1–8. [CrossRef] Diller, E.; Sitti, M. Three-dimensional programmable assembly by untethered magnetic robotic micro-grippers. Adv. Funct. Mater. 2014, 24, 4397–4404. [CrossRef] Servant, A.; Qiu, F.; Mazza, M.; Kostarelos, K.; Nelson, B.J. Controlled in vivo swimming of a swarm of bacteria-like microrobotic flagella. Adv. Mater. 2015, 27, 2981–2988. [CrossRef] [PubMed] Xu, T.; Yu, J.; Yan, X.; Choi, H.; Zhang, L. Magnetic actuation based motion control for microrobots: An overview. Micromachines 2015, 6, 1346–1364. [CrossRef] Kim, S.J.; Jeon, S.M.; Nam, J.K.; Jang, G.H. Closed-loop control of a self-positioning and rolling magnetic microrobot on 3D thin surfaces using biplane imaging. IEEE Trans. Magn. 2014, 50, 1–4. [CrossRef] Ghanbari, A.; Chang, P.H.; Nelson, B.J.; Choi, H. Magnetic actuation of a cylindrical microrobot using time-delay-estimation closed-loop control: Modeling and experiments. Smart Mater. Struct. 2014, 23, 35013–35024. [CrossRef]


22. 23. 24. 25. 26. 27.

14 of 14

Wong, D.; Steager, E.B.; Kumar, V. Independent control of identical magnetic robots in a plane. IEEE Robot. Autom. Lett. 2016, 1, 554–561. [CrossRef] Steager, E.B.; Sakar, M.S.; Magee, C.; Kennedy, M.; Cowley, A.; Kumar, V. Automated biomanipulation of single cells using magnetic microrobots. Int. J. Robot. Res. 2013, 32, 346–359. [CrossRef] Jiles, D.C. Introduction to Magnetism and Magnetic Materials; CRC Press: Boca Raton, FL, USA, 1998. Han, B.; Park, S.; Lee, S. Gradient waveform synthesis for magnetic propulsion using MRI gradient coils. Phys. Med. Biol. 2008, 53, 4639–4649. [CrossRef] [PubMed] Ocegueda, K.; Rodriguez, A. A simple method to calculate the signal-to-noise ratio of a circular-shaped coil for MRI. Concepts Magn. Reson. Part A 2006, 28, 422–429. [CrossRef] Yesin, K.B.; Vollmers, K.; Nelson, B.J. Modeling and control of untethered biomicrorobots in a fluidic environment using electromagnetic fields. Int. J. Robot. Res. 2006, 25, 527–536. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).