Portable manipulation system for commercial robotic surgery forceps

Transition from the utilisation of traditional instruments to new robotic methodologies in surgical operations occurs rapidly. Although the implementation of these methodologies to classical surgery operations is advantageous due to increased precisions and enhanced motion capabilities of robotic systems, overall worldwide accessibility and adaptation are still limited due to high equipment costs and special infrastructure requirements.

These systems usually consist of two sections as the manipulator itself performing macro manipulation and the multi degrees of freedom laparoscopic tool that performs micro manipulation inside the surgical target. Capability of macro-micro manipulation with redundant degrees of freedom (DoF >6) allows these robotic systems to perform challenging tasks at the surgery target by enhanced dexterity compared with limited degrees of freedom of classical laparoscopic instruments. 5,6 Addition of stereo vision systems to these surgical robots also eliminates the disadvantages of depth perception. Although the implementation of these systems is proved to be advantageous due to their enhanced operation capabilities, overall worldwide accessibility and adaptation are still limited due to high equipment costs, bulky sizes and special infrastructure requirements. In light of this, studies have been started to focus on the development of new laparoscopic instruments by increasing dexterities of classical ones in terms of possible articulations. The main idea behind this approach is to generate a compact and low-cost laparoscopic surgical instrument that will acquire robotic instrument's tip point manoeuvrability ( Figure 1A).
This new type of articulating laparoscopic instruments (ALI) can be described as classical laparoscopic instruments that are improved by the implementation of wrist mechanism to their end effectors.
Even though there are numerous types of wrist mechanisms in literature, they are mainly categorised according to their joint types.
The main difference between them is the articulation method of the end effector tip point. One of the most popular methods is to manipulate the end effector tip point around a pin joint, while the other one is utilising a compliant joint ( Figure 1B). 7,8 Advanced versions of ALI can also be either partially or fully actuated to form robotic articulating laparoscopic instruments (RALI).
The main advantages of RALI can be given as their enhanced motorised dexterity and additional adaptable flexibility in terms of the kinematics between the user input and the end effector tip point movement. On the other hand, while the main idea behind the motorisation of ALI is to improve dexterities of these tools, another important reason is the reduction of reaction forces formed between the device and body wall during manipulation due to enhanced motorised dexterity. 9 Therefore, due to RALI, more precise end effector tip point movement, application of reduced reaction forces to the body wall, and adaptive kinematic mapping between the user interface and wrist motion would be provided as these benefits are also provided by the features of expensive and complex robotic surgical systems. Throughout the literature, studies have been focused on the designs and operational verifications of RALI. Espiritu et al. 10 studied grasping and pinching force created by a RALI forceps jaw that is simulated on the pig tissue finite element method model. The results show that the deformation of pig tissue can occur easily in simulation trials. Jaimy 11 was developed as a RALI by having two degrees of freedom motorised jaws. The main advantage of Jaimy is the fact that it can be manipulated by a single finger only. Amato et al. 12 focused on the limited amount of force and torque application to a tissue. The authors have proposed an electro-mechanical system that limits force and torque delivered to the tissue using a laparoscopic instrument. In their design, pre-defined set points can be given by the surgeon as force/torque constraints. Focacci et al. 13 proposed a system having the actuation unit placed out of the main instrument to decrease overall weight. Actuator motions are transferred by utilising flexible components called as sheaths. Since the main objective of the authors' study is to propose a method for actuation transfer, they utilised commercial robotic forceps of daVinci system in their trials. Piccigallo et al. 14 developed a handheld lightweight and ergonomic robotic instrument with a 3 DoF roll pitch-roll end-effector that is actuated by the same method proposed by Focacci et al. 13 . Kim et al. 15 proposed a light weight portable surgical system called as S-Surge. In their work, the authors studied the robotic laparoscopic instrument that can execute both 3 DoF roll, wrist, grasping motions and 4 DoF remote centre of motion macro manipulator.
Considering given brief literature about ALI, it can easily be seen that the main focus is directed towards enhancing tool dexterity and DoF. On the other hand, to achieve that target, most of the proposed actuation systems have been integrated to the main body of the instruments. This condition, as a common disadvantage, renders sterilisation procedures complex and also causes conflicts about the instrument's amount of reusability for cost considerations. In light of this, the current paper focuses on the design of a modular and dexterous robotic articulating laparoscopic system to reach advantages of robotic surgical instruments within the compact forms of classical laparoscopic instruments. Throughout the study, the design of four degrees of freedom low cost, compact and portable manipulation system was proposed to adopt commercial robotic surgery forceps. In order to represent the main idea, this study has been carried out considering existing daVinci forceps. Having separate actuation units that allow easy attachment of robotic surgery forceps of daVinci system, the proposed instrument not only reduces reusability costs of the modular system but also provides a chance for health centres with insufficient infrastructure or budget to utilise the F I G U R E 1 (A) Classical and robotic laparoscopic forceps. (B) Forceps joint types. benefits of multi degrees of freedom robotic forceps. The main design goal of the proposed instrument is to provide robotic features by using daVinci robotic forceps as multi DoF laparoscopic tools and attachable lightweight actuation and controller box forming tool handle. The proposed design not only aims to provide cost effective system but also aims to provide adaptability in terms of the kinematic mapping between the user interface and forceps due to its embedded control algorithm. This way the surgeon could easily change scaling factors of the end effectors displacement/velocity and motion mappings.

| STRUCTURAL DESIGN
Earlier study of the authors 16 proved feasibility of the proposed concept by using simulation environments and low-resolution actuation prototypes. In order to propose a modular robotic articulating laparoscopic system that will adopt robotic surgery forceps and allow surgeons to use it comfortably as classical laparoscopic tools, this study firstly focuses on the ergonomic structural design of the system handling that will include separate modules as an actuation box to house actuators, forceps attachment slot, transmission couplings, controller joysticks/sensors, and an ergonomic handle for them to be installed. In light of this, design optimisation was carried out in a simulation environment by considering the previous study. 16 The multi-objective design optimisation in terms of reducing overall weight, improved ergonomics, and usability of physical human interface was carried out. Figure 2 illustrates the exploded model of the system with visible mentioned components.
Throughout the study, one of the most critical optimisation issues was taken as the overall weight of the system for improved ergonomics. During design iterations and prototype trials, it was observed that the vast majority of overall system weight contribution will be given by actuators. Thus, Dynamixel XL-430 smart actuators were selected to be used in the prototype for their relatively high torque to weight ratio. Following the selection of light weight actuators, distribution of the forces and moments that will affect system handling was considered. As seen in Figure 2, parameters a and b represent the dimensions between the actuation box centre of mass and the pivot point c of the manipulator. In order to improve handling, these parameters were designed by considering the weight F I G U R E 2 Exploded model of proposed system and handling dimensional parameters. YAZICI ET AL. of the actuators located at the top back of the system as they mostly contribute the overall system weight. Multiple low-resolution prototypes were manufactured and handled by the surgeons. Considering received feedbacks, these dimensions were adjusted to reduce generated moments at handling location to increase usability performance. Another step in the structural design of the proposed portable manipulation system is to provide attachment compatibility with daVinci robotic surgery forceps. Thus, the design of the system must fulfil the requirements of both geometrical and mechanical conditions. While geometric compatibility between the robotic forceps and proposed manipulation system can be fulfiled by a complex

| COUPLED MOTION ANALYSIS
The actuation of the selected robotic forceps is carried out by a tendon-driven structure; thus, there exists coupled motion formation generated by end effector pitch motion on yaw motion. Any angular displacement in the pitch axis generates a proportional angular displacement in the yaw axis. This behaviour results in an unwanted movement on the yaw axis. Unless eliminating this affect, surgeons should always try to compensate coupled motion during the operation.
Such an effort would make control of the forceps more complicated.
Therefore, in order to eliminate this effect automatically within the system software, a digital decoupling control algorithm was created.
The basic logic behind mentioned digital decoupling is the ability to provide a counter motion input in the opposite direction to compensate undesired coupled motion on the joint. Compensation amount was uniquely obtained from the measured dimensions of forceps itself. Considering the mechanical structure of forceps, the compensation multiplier as decoupling ratio (R d ) was defined to calculate the decoupling amount. The utilisation of this multiplier provides an input to the axis of the coupled motion proportional to the motion of pitch axis. Forceps behaviour affected by coupled motion is illustrated in Figure 4, where the resultant yaw axis motion is visible after an input applied to the pitch axis.
In order to calculate angle θ 3 that is formed undesirably by the actuation of pitch axis by θ 2 , the decoupling ratio should be derived uniquely for target robotic forceps (Equation 1).

| Derivation of the decoupling ratio (R d )
The main reason behind the coupled motion formation on daVinci robotic surgery forceps is the mechanical structure of the existing pulleys and wires. As the contact point of the wires and pulleys varies during forceps motion, there will be an additional motion formation on the joints that are actuated by these wires proportional to the amount of contact point variation. This phenomenon can be seen in  -YAZICI ET AL. S1 and S2 due to the forceps structure; these segment lengths stay constant during the forceps motion. On the other hand, length of the arc that is formed between P3 and P4 defining S3 segment varies by forceps motion. Since this difference causes coupled motion in the yaw axis, the above equation can also be written as If Equations (2) and (3) are combined together, the amount of coupled angular displacement on the yaw axis can be written as In order to compensate the motion, the required amount of input that should be applied by the system can be calculated as The relationship between θ 0 2 and driver knob angular displacement can also be similarly written as Combining Equations (5) and (6)

F I G U R E 4
Representation of undesired yaw axis motion by θ 3 after pitch axis input by θ 2 , daVinci forceps pulleys, and yaw axis cable route.

-5 of 11
Using Equation (7), R d can be calculated by using actual dimensions of forceps and driver pulley structures. Throughout the study, this ratio is calculated to be around the vicinity of R d ¼ 0:66 by measuring related parameters manually. In this scenario, the thumb finger will manipulate yaw and pitch motion by a two degrees of freedom joystick, the index finger will manipulate roll motion by a linear pressure sensor and finally the middle finger will manipulate forceps grasping motion by a simple pressure sensor (Figure 8).

| TEST PLATFORM DESIGN
The proposed portable manual laparoscopy manipulation system targets an enhanced dexterity through the surgical operations compared with classical laparoscopy tools. In order to verify this, a kind of test platform was decided to be designed ( Figure 9). As the proposed system has wider workspace with attached daVinci robotic forceps, the designed platform was utilised to carry out the hardware verification of the system. Similar to the proposed manipulation system, the structure of the test platform also has high modularity in terms of test scenarios. Therefore, the objects on the base matrix were designed in such a

F I G U R E 6
Form of the yaw axis transmission cable in two consecutive poses, qualitative illustration of S1 and S2 segment length preservation, and cable segments in consecutive forceps poses.
F I G U R E 7 Driver pulley knobs.
way that comparisons of both systems would be possible in terms end effector manoeuvrability. Figure 9 illustrates the main compo-

| PROTOTYPING AND CONTROL IMPLEMENTATION
In order to continue hardware verification tests, early prototypes of the proposed manipulation system and test platform were manufactured with the help of rapid prototyping ( Figure 10).
Following the assembly of electro-mechanical components, the proposed control methodology was implemented to the system ( Figure 11).
The proposed hardware and software should be integrated As mentioned in earlier sections, due to the tendon driven nature of the robotic forceps, the manipulation system is prone to coupled motions. In the regular basic control scheme, θ 2 input (Figure 8) from the handle will directly be transferred to the forceps pitch axis that creates an undesirable rotation on the yaw axis. If the user wants to compensate for this motion automatically, digital decoupling can be activated and θ 2 input from the handle will create a proportional pitch, and yaw motion on the forceps actuation. Table 1 shows the relation between sensory inputs and corresponding forceps motion.
Forceps motion under digital decoupling off and on is illustrated in Figure 12.
As seen in Figure 12, there exists a visible undesired yaw axis motion during the actuation of pitch axis via joystick input without any decoupling control. On the other hand, when decoupling is activated, the yaw axis motion is successfully compensated by the controller of the system by utilising the decoupled ratio.

| HARDWARE VERIFICATION
Although Figure 12 shows successful yaw axis compensation visibly during pitch up and pitch down motion, in order to refine the decoupling ratio and minimise measurement errors by verifying the smallest yaw axis deviation, a motion capture setup was prepared by utilising OptiTrack V100R2 motion capture cameras.
As seen in Figure 13, a single infra red reflector was attached to the forceps' end effector and dual motion capture cameras were utilised to track its motion. In order to stably fix the manufactured prototype and register the base plane, an attachment frame was prepared and assembled with the prototype. During verification trials, the same pitch up and down motion ( Figure 14) was executed on the forceps by using various compensation ratios (R d ) around the vicinity of calculated one.
F I G U R E 1 0 Prototypes of the manipulation system and test platform.
F I G U R E 1 1 Control methodology and hardware implementation.
However, prior to analysing yaw axis deviation results that are visible in the z axis efficiently, roll axis error due to the setup assembly should be corrected. Whether compensated or not, it is clear that during prescribed motion, the largest forceps tip travel (between max and min pitch motion) in the y axis occurs in corrected roll axis rotation. Using this information, roll axis correction angles and tip travel ranges were calculated for each of the yaw axis compensation ratios between 0.63 and 0.73. Following this procedure, the acquired data were plotted to reveal the yaw axis deviations that are visible in the z axis (Figure 15).
At the end of the experiment, square sums of each data for different compensation ratios were calculated and all f the acquired information is shown in Table 2.
As seen in Table 2  Kinematic mapping between actuators and forceps end effector was implemented by a control procedure that allows adaptability between the system and the surgeon in terms of modifiable scaling F I G U R E 1 5 Yaw axis deviations visible in the z axis. factors as displacement, velocity and motion mappings by feeding modified inputs to the actuators embedded position control. This control procedure not only allows surgeons to personalise the behaviour of the manipulation system but also compensates for the coupled motion of selected robotic forceps by the implementation of decoupling ratio. The decoupling ratio of the mechanism was firstly calculated by manually measuring knob/pulley diameters of the mechanism and found around the vicinity of 0.66. Later hardware verification studies show that the refined decoupling ratio of the utilised system is 0.70. Manual measurement errors can be accountable to the small difference between measured and experimented values. Throughout the study, the verification of the manufactured prototype was carried out on a proposed test setup and motion capture system. Although the current version of the prototype works in a tethered mode through a control PC, promising results were achieved in terms of increased dexterity and target reachability.