System and method for reducing interference in positional sensors for robotic surgery
The invention involves a system and method for increasing positional accuracy of surgical systems that utilize magnetic or electromagnetic sensors to provide positional awareness to a surgeon or robot performing the surgery. The system takes advantage of electromagnetic tracking through sensors. These sensors are very accurate and repeatable, while being compact enough to not inhibit surgical procedures. The accuracy and repeatability of the sensors is <1 mm within a predetermined 6 inch×6 inch performance motion box. The system is constructed and arranged to map the distortion patterns of the sensor and, in real time, correct the distortion pattern to provide accurate location of anatomical structures for performance of a surgery.
1 . A system for locating a position of a surgical tool held by a surgical robot with compensation for electromagnetic interference from the surgical tool, comprising:
a surgical tool for performing a medical procedure on a patient;
an electromagnetic sensor for sensing a position of the surgical tool;
a field generator adapted to generate a magnetic field around the electromagnetic sensor to induce current in the electromagnetic sensor;
a neural network which has been trained with distorted data representing a position of the electromagnetic sensor with interference from nearby objects, wherein the surgical robot is caused to move the surgical tool through a series of positions and each of the distorted data includes a current position and orientation of the surgical tool and distorted sensor data received by the electromagnetic sensor at the current position and orientation of the surgical tool, the neural network being trained until a threshold error reaches a predetermined threshold;
a control unit configured to feed a current electromagnetic sensor output to the trained neutral network and calculate the position of the electromagnetic sensor based on an output of the trained neutral network;
the surgical robot having an arm for holding the surgical tool, wherein the control unit is configured to receive the position of the electromagnetic sensor for positioning the robot arm based on the output of the control unit;
wherein the control unit is programmed to record the distorted data from the electromagnetic sensor for purposes of collecting data points for training the neural network;
wherein the control unit constructs a distortion map for the surgical tool, the distortion map representing a positional distortion between the electromagnetic sensor and the field generator for the surgical tool as the surgical tool is moved through the series of positions.
2 . The system of claim 1 , wherein the neural network comprises:
a first neural network which has been trained with distorted data representing a position of the electromagnetic sensor with interference from nearby objects; and
a second neural network which has been trained with distorted data representing an orientation of the electromagnetic sensor with interference from nearby objects.
3 . The system of claim 1 , wherein the distorted data for training the neural network includes a difference between distorted sensor data and undistorted sensor data.
4 . The system of claim 3 , wherein the difference between the distorted sensor data and the undistorted sensor data includes an X,Y,Z position and a rotation matrix representing an orientation of the electromagnetic sensor at the time of measurement.
5 . The system of claim 1 , wherein the distorted data used to train the neural network includes a current surgical tool position and orientation, a current electromagnetic transformation matrix and an electromagnetic sensor quality.
6 . The system of claim 1 , wherein the neural network has been iteratively trained until the difference between the distorted data and a predicted data is less than 0.01 mm and 0.01 degrees.
7 . The system of claim 1 , wherein the neural network has been trained with distorted data that has been normalized.
8 . The system of claim 1 wherein the robot is adapted to utilize two or more surgical tools, the control unit constructing a respective distortion map for each surgical tool.
9 . A system for locating a position of a surgical tool held by a surgical robot with compensation for electromagnetic interference, comprising:
a surgical tool for performing a medical procedure on a patient;
an electromagnetic sensor for sensing a location of the surgical tool;
a field generator adapted to generate a magnetic field around the electromagnetic sensor to induce current in the electromagnetic sensor;
a neural network which has been trained with distorted data representing a position of the electromagnetic sensor with interference from nearby objects, wherein the surgical robot is caused to move the surgical tool through a series of positions and each of the distorted data includes a current position and orientation of the surgical tool and distorted sensor data received by the electromagnetic sensor at the current position and orientation of the surgical tool, the neural network being trained until a threshold error reaches a predetermined threshold;
a control unit configured to feed a current electromagnetic sensor output to the trained neutral network and calculate the position of the electromagnetic sensor based on an output of the trained neutral network;
wherein the control unit is programmed to record the distorted data from the electromagnetic sensor for purposes of collecting data points for training the neural network;
wherein the control unit constructs a distortion map for the surgical tool, the distortion map representing a positional distortion between the electromagnetic sensor and the field generator for the surgical tool as the surgical tool is moved through the series of positions.
10 . The system of claim 9 , wherein the neural network comprises:
a first neural network which has been trained with distorted data representing a position of the electromagnetic sensor with interference from nearby objects; and
a second neural network which has been trained with distorted data representing an orientation of the electromagnetic sensor with interference from nearby objects.
11 . The system of claim 9 , wherein the distorted data for training the neural network includes a difference between distorted sensor data and undistorted sensor data.
12 . The system of claim 11 , wherein the difference between the distorted sensor data and the undistorted sensor data includes an X,Y,Z position and a rotation matrix representing an orientation of the electromagnetic sensor at the time of measurement.
13 . The system of claim 9 , wherein the distorted data used to train the neural network includes a current surgical tool position and orientation, a current electromagnetic transformation matrix and an electromagnetic sensor quality.
14 . The system of claim 9 , wherein the neural network has been iteratively trained until the difference between the distorted data and a predicted data is less than 0.01 mm and 0.01 degrees.
15 . The system of claim 9 , wherein the neural network has been trained with distorted data that has been normalized.
16 . The system of claim 9 wherein the robot is adapted to utilize two or more surgical tools, the control unit constructing a respective distortion map for each surgical tool.