Compatible Robot Research Articles

Dynamic path planning via Dueling Double Deep Q-Network (D3QN) with prioritized experience replay

Path planning is a key requirement for mobile robots employed for different tasks such as rescue or transport missions. Conventional methods such as A* or Dijkstra to tackle path planning problem need a premise map of the robot's environment. Nowadays, dynamic path planning is a popular research topic, which drives mobile robots without prior static requirements. Deep reinforcement learning (DRL), which is another popular research area, is being harnessed to solve dynamic path planning problem by the researchers. In this study, Deep Q-Networks, which is a subdomain of DRL are opted to solve dynamic path planning problem. We first employ well known techniques Double Deep Q-Networks (D2QN) and Dueling Double Deep Q-Networks (D3QN) to train a model which can drive a mobile robot in environments with static and dynamic obstacles within 3 different configurations. Then we propose D3QN with Prioritized Experience Replay (PER) extension in order to further optimize the DRL model. We created a test bed to measure the performance of the DRL models against 99 randomly generated goal locations. According to our experiments, D3QN-PER method performs better than D2QN and D3QN in terms of path length and travel time to the goal without any collisions. Robot Operating System and Gazebo simulation environment is utilized to realize the training and testing environments, thus, the trained DRL models can be deployed to any ROS compatible robot seamlessly.

Changes in Resting State Functional Connectivity Associated with Dynamic Adaptation of Wrist Movements.

Dynamic adaptation is an error-driven process of adjusting planned motor actions to changes in task dynamics (Shadmehr, 2017). Adapted motor plans are consolidated into memories that contribute to better performance on re-exposure. Consolidation begins within 15 min following training (Criscimagna-Hemminger and Shadmehr, 2008), and can be measured via changes in resting state functional connectivity (rsFC). For dynamic adaptation, rsFC has not been quantified on this timescale, nor has its relationship to adaptative behavior been established. We used a functional magnetic resonance imaging (fMRI)-compatible robot, the MR-SoftWrist (Erwin et al., 2017), to quantify rsFC specific to dynamic adaptation of wrist movements and subsequent memory formation in a mixed-sex cohort of human participants. We acquired fMRI during a motor execution and a dynamic adaptation task to localize brain networks of interest, and quantified rsFC within these networks in three 10-min windows occurring immediately before and after each task. The next day, we assessed behavioral retention. We used a mixed model of rsFC measured in each time window to identify changes in rsFC with task performance, and linear regression to identify the relationship between rsFC and behavior. Following the dynamic adaptation task, rsFC increased within the cortico-cerebellar network and decreased interhemispherically within the cortical sensorimotor network. Increases within the cortico-cerebellar network were specific to dynamic adaptation, as they were associated with behavioral measures of adaptation and retention, indicating that this network has a functional role in consolidation. Instead, decreases in rsFC within the cortical sensorimotor network were associated with motor control processes independent from adaptation and retention.SIGNIFICANCE STATEMENT Motor memory consolidation processes have been studied via functional magnetic resonance imaging (fMRI) by analyzing changes in resting state functional connectivity (rsFC) occurring more than 30 min after adaptation. However, it is unknown whether consolidation processes are detectable immediately (<15 min) following dynamic adaptation. We used an fMRI-compatible wrist robot to localize brain regions involved in dynamic adaptation in the cortico-thalamic-cerebellar (CTC) and cortical sensorimotor networks and quantified changes in rsFC within each network immediately after adaptation. Different patterns of change in rsFC were observed compared with studies conducted at longer latencies. Increases in rsFC in the cortico-cerebellar network were specific to adaptation and retention, while interhemispheric decreases in the cortical sensorimotor network were associated with alternate motor control processes but not with memory formation.

Review of Clinical and Technological Consideration for MRI-Guided Robotic Prostate Brachytherapy

Low Dose Rate Brachytherapy (LDR-BT) is a technique for treating localized prostate cancer by implanting radioactive seeds. In conventional practice, the delivery of seeds is performed using transrectal ultrasonography (TRUS) imaging for implant guidance and checked using computed-tomography for post-implant dosimetry. In the case of TRUS, accuracy can be compromised due to sub-optimal imaging. Magnetic Resonance Imaging (MRI), however, is known to provide better soft-tissue contrast, therefore, increasing the ability to detect small lesions; for that reason, the integration of intraoperative MRI in BT workflows has been investigated over the last two decades. The fusion of preoperative MR-images during TRUS-brachytherapy is possible. However, the image registration process introduces a source of uncertainty. Manual, real-time intra-operative LDR-BT is challenging under MRI due to confined space and procedural workflows. This motivates the development of MRI-compatible robots for prostate BT, with potential advantages of improved source placement accuracy and final dosimetry. In this paper, the state-of-art of technological components in MRI compatible robots, especially for LDR-BT, has been presented. This systematic review helps us to position an ongoing Cooperative Brachytherapy project, developing a real-time MRI-guided robot for adaptive LDR-BT. The design approach includes integrating separate modules: imaging, dose planning, needles, and robot.

Cortical Processing of Multimodal Sensory Learning in Human Neonates.

Following birth, infants must immediately process and rapidly adapt to the array of unknown sensory experiences associated with their new ex-utero environment. However, although it is known that unimodal stimuli induce activity in the corresponding primary sensory cortices of the newborn brain, it is unclear how multimodal stimuli are processed and integrated across modalities. The latter is essential for learning and understanding environmental contingencies through encoding relationships between sensory experiences; and ultimately likely subserves development of life-long skills such as speech and language. Here, for the first time, we map the intracerebral processing which underlies auditory-sensorimotor classical conditioning in a group of 13 neonates (median gestational age at birth: 38 weeks + 4 days, range: 32 weeks + 2 days to 41 weeks + 6 days; median postmenstrual age at scan: 40 weeks + 5 days, range: 38 weeks + 3 days to 42 weeks + 1 days) with blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (MRI) and magnetic resonance (MR) compatible robotics. We demonstrate that classical conditioning can induce crossmodal changes within putative unimodal sensory cortex even in the absence of its archetypal substrate. Our results also suggest that multimodal learning is associated with network wide activity within the conditioned neural system. These findings suggest that in early life, external multimodal sensory stimulation and integration shapes activity in the developing cortex and may influence its associated functional network architecture.

Development of Applications for Humanoid Robots Using Multiple Platforms, Tools, and Cloud Data Sharing

This paper describes the procedure followed for using third-party tools and applications, avoiding the development of complex communication software modules for data sharing. A common practice in robotics is the use of middlewares to interconnect different software applications, hardware components, or even complete systems. It allows code and tool reuse minimizing the development effort. In this way, applications developed for one middleware can be shared with others by means of establishing communication bridges among them. The most extended procedure is the development of software modules that use the low-level communication resources that middlewares provide. This procedure has many advantages but a clear disadvantage: the complexity of development. The procedure proposed is based on the use of cloud technologies for data sharing without the development of middleware bridges. The way of inter-relate different middlewares is by means of the development of a compatible robot model. This procedure has enabled the use of the ArmarX middleware tools and the application of the results obtained to the humanoid robot TEO, that uses the YARP middleware, in an easy and fast way.

Pressure Observer Based Adaptive Dynamic Surface Control of Pneumatic Actuator with Long Transmission Lines

In this paper, the needle insertion motion control of a magnetic resonance imaging (MRI) compatible robot, which is actuated by a pneumatic cylinder with long transmission lines, is considered and a pressure observer based adaptive dynamic surface controller is proposed. The long transmission line is assumed to be an intermediate chamber connected between the control valve and the actuator in series, and a nonlinear first order system model is constructed to characterize the pressure losses and time delay brought by it. Due to the fact that MRI-compatible pressure sensors are not commercially available, a globally stable pressure observer is employed to estimate the chamber pressure. Based on the model of the long transmission line and the pressure observer, an adaptive dynamic surface controller is further designed by using the dynamic surface control technique. Compared to the traditional backstepping design method, the proposed controller can avoid the problem of “explosion of complexity” since the repeated differentiation of virtual controls is no longer required. The stability of the closed-loop system is analytically proven by employing the Lyapunov theory. Extensive experimental results are presented to demonstrate the effectiveness and the performance robustness of the proposed controller.

Robotically Assisted Long Bone Biopsy Under MRI Imaging: Workflow and Preclinical Study

Our research team has developed a magnetic resonance imaging (MRI)-compatible robot for long bone biopsy. The robot is intended to enable a new workflow for bone biopsy in pediatrics under MRI imaging. Our long-term objectives are to minimize trauma and eliminate radiation exposure when diagnosing children with bone cancers and bone infections. This article presents our robotic systems, phantom accuracy studies, and workflow analysis. This section describes several aspects of our work including the envisioned clinical workflow, the MRI-compatible robot, and the experimental setup. The workflow consists of five steps and is intended to enable the entire procedure to be completed in the MRI suite. The MRI-compatible robot is MR Safe, has 3 degrees of freedom, and a remote center of motion mechanism for orienting a needle guide. The accuracy study was done in a Siemens Aera 1.5T scanner with a long bone phantom. Four targeting holes were drilled in the phantom. Each target was approached twice at slightly oblique angles using the robot needle guide for a total of eight attempts. A workflow analysis showed the average time for each targeting attempt was 32 minutes, including robot setup time. The average 3D targeting error was 1.39 mm with a standard deviation of 0.40 mm. All of the targets were successfully reached. The results showed the ability of the robotic system in assisting the radiologist to precisely target a bone phantom in the MRI environment. The robot system has several potential advantages for clinical application, including the ability to work at the MRI isocenter and serve as a steady and precise guide.

Towards a compact and precise sample holder for macromolecular crystallography.

Most of the sample holders currently used in macromolecular crystallography offer limited storage density and poor initial crystal-positioning precision upon mounting on a goniometer. This has now become a limiting factor at high-throughput beamlines, where data collection can be performed in a matter of seconds. Furthermore, this lack of precision limits the potential benefits emerging from automated harvesting systems that could provide crystal-position information which would further enhance alignment at beamlines. This situation provided the motivation for the development of a compact and precise sample holder with corresponding pucks, handling tools and robotic transfer protocols. The development process included four main phases: design, prototype manufacture, testing with a robotic sample changer and validation under real conditions on a beamline. Two sample-holder designs are proposed: NewPin and miniSPINE. They share the same robot gripper and allow the storage of 36 sample holders in uni-puck footprint-style pucks, which represents 252 samples in a dry-shipping dewar commonly used in the field. The pucks are identified with human- and machine-readable codes, as well as with radio-frequency identification (RFID) tags. NewPin offers a crystal-repositioning precision of up to 10 µm but requires a specific goniometer socket. The storage density could reach 64 samples using a special puck designed for fully robotic handling. miniSPINE is less precise but uses a goniometer mount compatible with the current SPINE standard. miniSPINE is proposed for the first implementation of the new standard, since it is easier to integrate at beamlines. An upgraded version of the SPINE sample holder with a corresponding puck named SPINEplus is also proposed in order to offer a homogenous and interoperable system. The project involved several European synchrotrons and industrial companies in the fields of consumables and sample-changer robotics. Manual handling of miniSPINE was tested at different institutes using evaluation kits, and pilot beamlines are being equipped with compatible robotics for large-scale evaluation. A companion paper describes a new sample changer FlexED8 (Papp et al., 2017, Acta Cryst., D73, 841-851).

Design and analysis of a tendon-based computed tomography-compatible robot with remote center of motion for lung biopsy.

Nowadays, biopsy is a decisive method of lung cancer diagnosis, whereas lung biopsy is time-consuming, complex and inaccurate. So a computed tomography-compatible robot for rapid and precise lung biopsy is developed in this article. According to the actual operation process, the robot is divided into two modules: 4-degree-of-freedom position module for location of puncture point is appropriate for patient's almost all positions and 3-degree-of-freedom tendon-based orientation module with remote center of motion is compact and computed tomography-compatible to orientate and insert needle automatically inside computed tomography bore. The workspace of the robot surrounds patient's thorax, and the needle tip forms a cone under patient's skin. A new error model of the robot based on screw theory is proposed in view of structure error and actuation error, which are regarded as screw motions. Simulation is carried out to verify the precision of the error model contrasted with compensation via inverse kinematics. The results of insertion experiment on specific phantom prove the feasibility of the robot with mean error of 1.373 mm in laboratory environment, which is accurate enough to replace manual operation.

Development of MRI Compatible Tapping Assistive Robot

Development of MRI Compatible Tapping Assistive Robot

Investigation of active tracking for robotic arm assisted magnetic resonance guided focused ultrasound ablation.

Focused ultrasound surgery (FUS) is a technique that does not need invasive access to the patient while allowing precise targeted therapy. Magnetic resonance (MR) guided FUS provides capabilities for monitoring treatments. Because the targeted tumours are distributed at different positions, focus repositioning becomes necessary. An MR compatible robot was used to increase the operational range of FUS application. Active tracking was developed to detect the robotic arm with regard to the MR coordinate system. The purpose of this study was to construct active tracking to allow a wide spatial range of repositioning the FUS transducer that is fast and accurate. The technique was characterised and validated by a series of positioning tests to prove its efficiency for guiding the robot. In the calibration range, tracking achieved an RMS accuracy of 0.63 mm. Results of phantom ablation showed a focal scanning precision Δx = 0.4 ± 0.37 mm, Δy = 0.4 ± 0.28 mm and Δz = 0.7 ± 0.66 mm. The active tracking localisation can be considered a feasible approach for the MR guided FUS system positioned by a robot. Copyright © 2016 John Wiley & Sons, Ltd.

Accurate error compensation for a MR-compatible surgical robot based on a novel kinematic calibration method

Kinematic calibration is an effective and economical way to improve the accuracy of surgical robot, and in most cases, it is a necessary procedure before the robot is put into operation. This study investigates a novel kinematic calibration method where the effect of controller error is taken into account when formulating the model based on screw theory, which is applied to the kinematic control of magnetic resonance compatible surgical robot. Based on screw theory, the kinematic error model is established for the relationship between error of controller and the deviation of the measured pose of the end-effector. Therefore, the error of controller can be figured out and parameters of controller can be adjusted accordingly. Control strategy based on the kinematic calibration framework is proposed. According to artificial neural network, the deviation of end-effector in arbitrary configuration can be effectively obtained. Comparative experiments are carried out to show the validity and effectiveness of the proposed framework with the help of commercial visual system and joint encoders.

On the Modulation of Brain Activation During Simulated Weight Bearing in Supine Gait-Like Stepping.

To date, the neurophysiological correlates of muscle activation required for weight bearing during walking are poorly understood although, a supraspinal involvement has been discussed in the literature for many years. The present study investigates the effect of simulated ground reaction forces (0, 20, and 40 % of individual body weight) on brain activation in sixteen healthy participants. A magnetic resonance compatible robot was applied to render three different levels of load against the feet of the participants during active and passive gait-like stepping movements. Brain activation was analyzed by the means of voxel-wise whole brain analysis as well as by a region-of-interest analysis. A significant modulation of brain activation in sensorimotor areas by the load level could neither be demonstrated during active nor during passive stepping. These observations suggest that the regulation of muscle activation under different weight-bearing conditions during stepping occurs at the level of spinal circuitry or the brainstem rather than at the supraspinal level.

Modeling and control of a high-precision tendon-based magnetic resonance imaging–compatible surgical robot

This article mainly includes the developing, dynamical modeling and control of a tendon-based robot system. First, a 5-degree-of-freedom tendon-based magnetic resonance imaging–compatible robot for prostate needle insertion surgery is introduced briefly. What follows is the dynamical modeling of the robot system, where a mechanical dynamic model is established using the Lagrange method, and a lumped parameter tendon model is used to identify the nonlinear gain of the actuator. Based on the dynamical model, a fuzzy sliding mode control algorithm is proposed for accurate position control of the robot. Through simulations using different sinusoidal input signals, we observed that the sinusoidal tracking error at 1/2π Hz is 0.2 mm and the needle tip positional precision of tracking a spatial arched curve remains less than 0.3 mm. Finally, experiments on tendon-sheath transmission and robot position tracking are conducted, which shows that the insertion precision is 0.67 mm in laboratory environment.

Design of a Dedicated Five Degree-of-Freedom Magnetic Resonance Imaging Compatible Robot for Image Guided Prostate Biopsy

In order to improve the current clinical application of magnetic resonance (MR)-guided prostate biopsies, a new, fully magnetic resonance imaging (MRI)-compatible solution has been developed. This solution consists of a five degree-of-freedom (5DOF) pneumatic robot, a programmable logic controller (PLC), and a software application for visualization and robot control. The robot can be freely positioned on the MR table. For the calibration of the robot and MR coordinate system, the robot’s needle guide (NG) is used. The software application supports the calibration with image segmentation and graphic overlays and guides the user through the interventional planning process. After selecting a target point, the application calculates the needed movements via solving the kinematics of the robot and translating the adjustment into commands for the PLC driving the step motors of the robot. In case further adjustments are required, the software also allows for manual control of the robot, to position the NG according to the acquired MR images.

A low-cost, high-field-strength magnetic resonance imaging-compatible actuator.

To perform minimally invasive surgical interventions with the aid of robotic systems within a magnetic resonance imaging scanner offers significant advantages compared to conventional surgery. However, despite the numerous exciting potential applications of this technology, the introduction of magnetic resonance imaging-compatible robotics has been hampered by safety, reliability and cost concerns: the robots should not be attracted by the strong magnetic field of the scanner and should operate reliably in the field without causing distortion to the scan data. Development of non-conventional sensors and/or actuators is thus required to meet these strict operational and safety requirements. These demands commonly result in expensive actuators, which mean that cost effectiveness remains a major challenge for such robotic systems. This work presents a low-cost, high-field-strength magnetic resonance imaging-compatible actuator: a pneumatic stepper motor which is controllable in open loop or closed loop, along with a rotary encoder, both fully manufactured in plastic, which are shown to perform reliably via a set of in vitro trials while generating negligible artifacts when imaged within a standard clinical scanner.

Assessment of MR compatibility and machinability using the fuzzy comprehensive evaluation method for material selection of MR device

ABSTRACTObjectiveTo develop magnetic resonance (MR) robot used in the magnetic field environment, materials selected for manufacturing robot should have good MR compatibility and machinability. Therefore, it is becoming increasingly urgent to develop an evaluation method considering the above two factors at the same time.MethodsA series of experiments were done to study the interaction between the magnetic field of MR scanner and the material samples, and to determine the samples’ machinability. Four kinds of materials including metal and plastics with different geometric shape were used to test the artifacts, force and torque in magnetic resonance image (MRI) tube according to standard method. Three items including surface roughness, cutting force and chip shape were analyzed to evaluate the machinability of four samples. The fuzzy comprehensive evaluation method was adopted to process these experimental data for choosing the most suitable robot material in MR environment.ResultsThe quantized values of MR compatibility and machinability are achieved, and a new methodology for material selection are developed, which provides systematic guideline for engineer and researcher to choose suitable robot materials used in an MRI tube. Polyformaldehyde of four kinds of material is determined to be the most suitable one for MR robots or devices.ConclusionA thorough examination of four commonly used engineering materials has been discussed and the fuzzy comprehensive evaluation method is suitable for choosing MR compatible robot materials. © 2015 Wiley Periodicals, Inc. Concepts Magn Reson Part B (Magn Reson Engineering) 45B: 1–12, 2015

The Application of Medical Robotics in the MRI-Guided Invention

This proposal focuses on the interventional MRI technology and the MRI compatible robots which are used to overcome some limitation of the MRI. In the introduction part, the development and the limitation of the current MRI interventions are introduced and a possible solution, MRI compatible robotic assistance, is proposed. In the design part, the specification of the MRI robot is discussed, including the MRI compatible material, the motor and encoder, the control unit and the combination of haptic sensor. Finally, the possible future application of the MRI compatible interventional robots is introduced.

Tracking of MR compatible interventional robots by controlling the MRI visibility of optically detunable MR markers

Tracking of MR compatible interventional robots by controlling the MRI visibility of optically detunable MR markers

Density-Convex Model Based Robust Optimization to Key Components of Surgical Robot

This paper investigates a new robust optimization framework based on density-convex reliability model and applies it to the dimensional optimization of magnetic resonance (MR) compatible surgical robot. As a justified tool for assessing reliability, the density-convex model is proposed on account of the reality that available data information is always insufficient. Based on the density-convex model, reliability functions of structure are constructed and taken as constraint conditions. The Euclidean norm of the sensitivity Jacobian matrix is selected as robust index and stated as the ultimate objective function. By using finite element method and artificial neural network (FEM–ANN) method, the explicit functions of mechanical response are achieved effectively. The optimization is solved by a gradient-based optimization algorithm in the framework. As an application of the above optimization framework, a prototype robot is designed and manufactured. Finally, a test experiment verifies the high reliability of the robot and further proves the validity and effectiveness of this proposed method.