WP3 Robotic manipulators in unstructured space
The primary goal of WP3 is to implement an automated procedure to define the position and orientation of the main components of the FUTURA platform with respect to a common reference frame; the aim is to implement a procedure with an effective time saving compared with current techniques. Furthermore, it will be essential to implement collision avoidance algorithms, between the two robots and with obstacles, both during off-line programming and on-line program execution. A safety exit strategy during on-line execution will also be defined. A planning/simulation/visualization software will be exploited for preliminary assessment of the proposed procedures. Further steps will involve the simulation of a real scenario, by means of the introduction in the platform workspace of artificial objects, also during the simulation of a real procedure.
DESCRIPTION OF WORK
T3.1 Robotic platform registration - Aim of Task activity is to aptly register within the operative room the overall robotic platform, i.e. the two robots, two probes, focused ultrasound surgery transducer and surgical table; this is highly relevant since the robot must own the information about its and the surrounding 3D environment relative and absolute position. This Task will be achieved thanks to several sensors, e.g. optic and/or magnetic; these sensors will be embedded on both the robotic platform and in the operative rooms, making it a smart-sensorized space.
T3.2 Collision avoidance between two robotic manipulators - It aims at providing collision avoidance algorithm both for positioning /planning phase (off-line) and for therapy deliver (on-line). Depending which phase we are considering, it must be ensured that while a robot moves freely to reach its goal position (target space for imaging robot or target organ for therapy robot) the second maintains its position avoiding collision with the other robot. This can be done more or less automatically depending on the kinematic redundancy respect to the planned task and robot workspace. In case the automatic algorithm fails, the user must have the possibility to re-plan the operation in order to fix the problem. A proper planning/simulation/visualization software will be developed to this purpose, trying to make it as easy as possible. In a second phase, during on-line execution, appropriate real-time algorithm will manage any difference with the planned scenario, adding real-time correction to the robot
motion in order to react to moving or unplanned obstacles or to adjust treatment path based on image guidance. Since the robots will not follow the planned path, also collision between robots must be avoided again. If the real time algorithm fails, robots will be halted and the system will prepare to the appropriate safety-exit strategy.
T3.3 Obstacles perception and avoidance - Basing on sensor information on the operating room (which will provide a 3D vision of the entire environment), robots must avoid collision with static and moving obstacles (other surgical instruments, physician, patient). This is performed in two distinct phases: during off-line programming, and during on-line execution, each with a dedicated algorithm. In the off-line programming phase, if the automatic algorithm fails, a proper planning/simulation/visualization software will allow re-planning and fixing by the user, in the easiest possible way. In the on-line execution phase, beyond the obstacle avoidance, the algorithm will drive the safety mechanism to slow down the robot velocity in case of hazardous situations and ultimately to halt the robot and prepare it to the appropriate safety-exit strategy.
T3.4 Implementation of a safety strategy (SM, SSSA) While safety issue will be addressed since the beginning of the project, they will be implemented as the platform development reaches an advanced stages during the on-line execution phase, a safety-exit strategy planning is continuously computed and updated, in order to have a safe response to unpredicted situations. Basely, after a controlled stop, the robots must be able to disengage themselves from the present obstacles and to reach a safe park position.