Robot arm modelling

A system for accurately determining a shape of a flexible arm robot, the system comprising:

• a flexible arm robot having a distributed Bragg reflector integrated into the arm, with the flexible robotic arm being coupled to an actuator pack, which in turn is connected to a computer system with software for controlling the robot, and the distributed Bragg reflector being connected to an interrogator measuring a change in Bragg wavelength as the flexible arm robot bends, and wherein the flexible arm robot has at least two tracking markers positioned on an outside surface thereof;
• at least two sensors used to track and record a movement and deformation of the flexible arm robot, and
• the computer system is connected to the interrogator and the at least two sensors, the computer system having a machine learning algorithm loaded onto the computer system, the machine learning algorithm being used to determine a calibration function for the position of the distributed Bragg reflector;
• wherein the at least two sensors are angularly and spatially distanced from each other, and
• wherein the machine learning algorithm is configured to optimize and match a sensor tracked backbone curve by performing a local/global optimization algorithm that rotates a sensing fiber backbone curve in X, Y, and Z-planes across 360 degrees to find a closest match to a sensor measured curve.