During April-May 2013, initial integration of the US system and Robotic system was performed. We received the latest version of the KUKA LWR, and are arranging to send the previous one back for replacement/repairs. After setting up the new robot, tests were performed on an artificial bone as seen in the photo. The 2D US probe was manipulated freehand, and the drill was co-manipulated with the robot arm. Testing was performed to assess the best free-hand motion for the probe, and the achievable error in the position of the drilling point and orientation.

Preliminary results showed errors below 10mm, highly dependent of the free-hand motion of the probe. Without the US system, and attaching the optical markers directly to the bone, the achievable error is in average 1mm in position and .01rad in orientation. Further testing in continuing, in order to lower the error with the US system, as well as the registration time cycle. 

Currently the robotic system is fully setup. Since the last report, the Polaris optical system has been added. The optical system is being used to assess the LWR precision, as well as its calibration. Optical frames have been attached to the robot end-effector and the bone. Calibration is performed through pivoting motions of the optical frames for both the bone and robot end-effector. The main focus is to quantify maximum error in relative positioning and orientation between the drill and the bone. This maximum error is using the native cartesian impedance controller of the LWR.

The 2D ultrasound system is now being prepared for integration with the robot arm. An artificial bone embedded in silicone gel is being prepared, which will serve for the first tests. Three optical frames are initially used, one on the bone, one on the ultrasound probe, and one on the dril (robot end-efffector). Cross comparisons will allow us to fully assess precision under impedance control. The first system integration is planned for the last week of April.

At this time the developed ultrasound based navigation system is capable to perform: 

    - 2D ultrasound based bone tracking, 

    - visualization of 3D bone point clouds, 

    - global registration of intra-operative to pre-operative bone point clouds,

- Novel algorithms were developed based on the ROS/OROCOS/PCL open source software tools, in the previous 3 modules of the software.

- The results obtained produced the paper "On the development and simulation of a Robotic Ultrasound Guided System for Orthopedic Surgical Procedures", which was submitted to the conference Robotica2013.

- The 2D ultrasound based bone tracking module has a mean error of 1.4 mm.

- The registration modules have a mean error of 1.27 mm and 1.63 mm for the initial calibration and the intra-operative registration, respectively.

- A simulation setup for the overall system is under development for the MORSE simulator.

- First steps towards an ontology for the Orthopedic Surgery domain were submitted to Robotics and Autonomous Systems journal "Ontology Applications and Standards in Service Robotics", a joint work with the IEEE Standard for Ontologies for Robotics and Automation, working group.

The robotic system setup is shown in the figure below. The robotic arm may be co-manipulated or tele-manipulated in order to correctly position the drill over the femural head in the simulator. 

The ultrasound based navigation system consisting in the extraction of intraoperative 3D ultrasound bone surfaces from ultrasound images, obtained without incisions in both free‑hand and robotized setups is now implemented and fully functional. In this time period robustness tests and development of alternative segmentation procedures have been tackled, results are expected in the next months.

- Extensive tests are going on for the "2D ultrasound based bone tracking" and "global registration of intra-operative to pre-operative bone point clouds", both core modules of the system.

- The team is following the works of the ISO/IEC “Joint Work Group on Standard for Medical Robots” and also actively contributing to the development of an IEEE Standard for Ontologies for Robotics and Automation, namely in the Service Robot Field (Orthopedic Surgery).

The HipRob experiment is going according to plan.

Both the co-manipulation robot controller and the ultrasound system are working and are now being prepared for integration

Robot Controller:

For the robot controller both co-manipulation and tele-manipulation have been implemented. The tele-manipulation may be very useful in a teaching setup where a more experienced surgeon may guide, through a haptic interface, a less experienced surgeon co-manipulating the surgical robot. This new feature has come up from the discussion meetings between the medical doctors and the engineers. A virtual surgical system has been implemented in CHAI3D, and is being used for development.

Ultrasound system:

The ultrasound system using 2D probes has evolved as initially planned and been implemented as last reported. Although the set-up is still relatively well structured, registration of a small US point cloud onto the CT point cloud has been achieved at aproximately 2 seconds. As we release the structure constraints, it is expected the registration times will increase. We are working on finding the optimal set-up which is both functional and real-time compatible for the surgical procedure.