Filip Šuligoj

Medical applicability of a low-cost industrial robot arm guided with an optical tracking system

Robot systems used in surgical procedures can autonomously position tools at points correlated with preoperative imaging techniques such as magnetic resonance (MR) and computed tomography (CT). The aim of this paper is to measure and assess medical applicability of a low-cost, lightweight industrial robot arm (Universal robot UR5) guided with the medically certified optical tracking system (Polaris Vicra) to positions registered from a CT scan. Technical setup, measurement equipment, device communication and robot control based on OTS feedback are described. Robot intrinsic accuracy, CT scan accuracy and two methods of robot tool positioning with aid of the optical tracking system (OTS) are measured. Measurements show RMS error of the robot (0.669 mm) is decreased 55.4% when guided with OTS using a single marker probe (0.29 mm) and 40.5% when using OTS with relative referencing (0.39 mm). RMS error of the CT scan readings is 0.46 mm.

Brain biopsy performed with the RONNA G3 system: a case study on using a novel robotic navigation device for stereotactic neurosurgery

Robotic neuronavigation is becoming an important tool for neurosurgeons. We present a case study of a frameless stereotactic biopsy guided by the RONNA G3 robotic neuronavigation system.

Validation of Three KUKA Agilus Robots for Application in Neurosurgery

In this paper, we verify three different 6 degrees of freedom Kuka Agilus robots for application in neurosurgery. Application specific reachability maps are generated for robots with 707 mm (R700), 901 mm (R900), and 1101 mm (R1100) horizontal reach. The reachability of each robot reflects a working volume of a standard stereotactic frame which utilizes the center of arc principle. A working volume with 100% reachability yield has been identified for the R900 and R1100 robots when the robot is positioned sideways to the patient. The R700 robot doesn’t have a 100% reachability yield work volume. Robot configurations within the reachability map are further optimized given two dexterity performance indices: the condition number and a new fuzzy joint limit avoidance function. In the experiments, we have further evaluated the impact on robot work volume given robot orientation with respect to the patient. After reorienting the robot a significant increase in work volume with 100% reachability yield was obtained for all three robots.

Automated Marker Localization in the Planning Phase of Robotic Neurosurgery

Accurate patient registration is a critical issue in medical image-guided interventions. The neurosurgical robotic system RObotic Neuro-NAvigation (RONNA) uses four retro-reflective spheres, on a marker attached to the patient’s cranial bone, for patient registration in physical and image space. In this paper, the algorithm for automatic localization of spherical fiducials in CT scans is presented and clinically evaluated. The developed localization algorithm uses a unique approach, which combines machine vision algorithms, biomedical image filtration methods, and mathematical estimation methods. The performance of the localization algorithm was evaluated in comparison with four skilled human operators. The measurements were based on twelve patient and eight lab phantom CT scans. The localization error of the algorithm in comparison with the human readings was smaller by 49.29% according to the ground truth estimation and by 45.91% according to the intra-modal estimation. Localization processing time was reduced by 84.96%. Reliability in terms of successful localization of the fiducial marker was 100% for 20 different test samples containing a total of 116 spherical fiducials. Based on the tests carried out in clinical conditions, the localization algorithm has demonstrated reliability with a high degree of accuracy and short processing time. The developed algorithm provides fully automated and accurate machine vision-based patient localization for the neurosurgical clinical application of the robotic system RONNA.

ARTgrid: a two-level learning architecture based on adaptive resonance theory

This paper proposes a novel neural network architecture based on adaptive resonance theory (ART) called ARTgrid that can perform both online and offline clustering of 2D object structures. The main novelty of the proposed architecture is a two-level categorization and search mechanism that can enhance computation speed while maintaining high performance in cases of higher vigilance values. ARTgrid is developed for specific robotic applications for work in unstructured environments with diverse work objects. For that reason simulations are conducted on random generated data which represents actual manipulation objects, that is, their respective 2D structures. ARTgrid verification is done through comparison in clustering speed with the fuzzy ART algorithm and Adaptive Fuzzy Shadow (AFS) network. Simulation results show that by applying higher vigilance values (ρ > 0.85) clustering performance of ARTgrid is considerably better, while lower vigilance values produce comparable results with the original fuzzy ART algorithm.

A Reinforcement Learning Based Algorithm for Robot Action Planning

The learning process that arises in response to the visual perception of the environment is the starting point for numerous research in the field of applied and cognitive robotics. In this research, we propose a reinforcement learning based action planning algorithm for the assembly of spatial structures with an autonomous robot in an unstructured environment. We have developed an algorithm based on temporal difference learning using linear base functions for the approximation of the state-value-function because of a large number of discrete states that the autonomous robot can encounter. The aim is to find the optimal sequence of actions that the agent (robot) needs to take in order to move objects in a 2D environment until they reach the predefined target state. The algorithm is divided into two parts. In the first part, the goal is to learn the parameters in order to properly approximate the Q function. In the second part of the algorithm, the obtained parameters are used to define the sequence of actions for a UR3 robot arm. We present a preliminary validation of the algorithm in an experimental laboratory scenario.

The Case of Industrial Robotics in Croatia

This paper presents an analysis of the number and the distribution of industrial robots in the Republic of Croatia. Also, the actual state of industrial robotics in the world is given, with the present and future growth trends, the distribution of robots by countries and manufacturing sectors. The number of robots in Croatia was obtained on the basis of a survey questionnaire sent to 1,500 Croatian companies. Regarding the question of robot ownership, 72 companies answered positively, resulting in a total of 326 active industrial robots in 2017. According to the Croatian Chamber of Economy estimates, the Croatian economy should have at least 2,000 installed robots. The paper gives a prediction for the growth trend in order to achieve 2,000 robots in a reasonable time period (by 2026). For that case, an exponential growth rate of 25.4% is required. Based on the current state of the Croatian economy, such an exponential growth is a huge challenge for the near future. The paper gives a brief critical review of the current state of industrial robotics in Croatia and provides guidelines for stimulating the application of industrial robots in the near future.

Position planning for collaborating robots and its application in neurosurgery

Primjena robotskih manipulatora u medicini danas je vrlo aktualno podrucje istraživanja. Unatoc tome jos uvijek postoji velik broj problema koji se javljaju kod pripreme vecine robotiziranih operacijskih postupaka. Jedan od glavnih je pozicioniranje robota u odnosu na pacijenta. Kod postavljanja robota u odnosu na unaprijed poznate radne tocke potrebno je osigurati efikasnu poziciju robota iz koje se sve zadane kretnje mogu izvrsiti bez kinematskih problema i kolizija. U radu je predstavljena metoda za planiranje prostornog razmjestaja robota prikladna za primjenu u neurokirurgiji. Razvijena metoda pociva na viseciljnoj optimizaciji funkcije cilja koja je sastavljena od kriterija koji objedinjuju prostornu upravljivost robota sa izbjegavanjem kolizija. Primjena razvijene metode validirana je na dvorucnom sustavu robota.