Device Technologies and Biomedical Robotics
Madeline Dudley (she/her/hers)
Biomedical Engineering Undergraduate Student
North Carolina State University
Raleigh, North Carolina, United States
Philip Brown
Professor
Wake Forest University/Virginia Tech SBES, United States
3D bioprinting is a novel method to produce 3D living structures from biomaterials that can be used in many biomedical applications such as healing wounds or repairing defects. Currently, the most common method of 3D bioprinting uses desktop 3D printers because of their high resolution and print quality. Although bioprinting has significantly advanced the field of tissue engineering, it faces many challenges in clinical translation. Some of these challenges include structural damage to the print when transferring it from the printer to the patient, high contamination risks, and very strict sterile requirements.1 Most importantly, this technology takes a significant amount of time ( >2 months). Real-time in-situ bioprinting aims to deposit biomaterials during surgery directly in the patient with the assistance of a robotic surgery system. Contrary to the flat printing surfaces in bioprinting using 3D printers, physiological structures have texturized surfaces that are not flat. This will result in variable printing distance; the distance between the printing nozzle and the substrate; if the tissue texture is not compensated for. Literature has shown that printing distance is critical to the quality and adhesion of the print. Excessive distance resulted in thin strands of resin that would not adhere appropriately while insufficient distance resulted in a blobbing effect. In both cases, the resultant print did not match the engineered structure. In this study, we develop an adaptive topology tracing algorithm that is capable of compensating for substrate texture, keeping the printing distance within the acceptable range for optimal print results.
A compensation and extruding system was developed to rigidly couple to the end effector of a da Vinci classic robot tool. The system is composed of the following: A Micro-Epsilon confocal chromatic sensor that is used to scan the distance from the extruder to the target surface. This sensor was chosen for its size, high resolution, and measurement range. Piezoelectric actuation systems were used to control the distance between the printing nozzle and the substrate and drive the piston of the extruder. These motors were chosen for their low weight, quick acceleration, and high resolution. The experimental setup is as follows: The system discussed earlier is held above custom-made randomly texturized silicone discs that simulate physiological tissue to be used as printing surfaces. These discs rigidly attach to velocity-controlled rotary motors. A graphical user interface was created in Python to control and monitor the developed system, control the rotary motor connected to the discs, and collect and plot real-time data from the motor’s encoder and the distance sensor. An adaptive velocity control algorithm was developed to manipulate the commands to the compensating motor in real time based on the input from the confocal sensor. The algorithm considers inherent system parameters such as natural and computational latency. Experiments were run to test how well the system could follow the texturized discs' profiles while they were rotating at different velocities. The root mean square error (RMSE) between the sensor’s measurements and the motor’s encoder was used as the error metric.
Results and Discussion: Current results show that the compensatory motion was able to follow the sensor input with an RMSE of 0.002 mm. This is within the acceptable bioprinting range of 50 microns. We observe that the compensatory motion can sometimes overshoot the target. This is possibly due to the current control methods and controller parameters. The control algorithm is continuously being developed and tuned to enhance performance. We also observed that increasing the linear velocity of the rotary motor resulted in an increased RMSE. We hypothesize that this is related to the natural and computational latency of the algorithm. The maximum velocity, which is currently 75 mm/s, that we are able to compensate for within acceptable error bands will be a limitation of the system.
Conclusion: The results demonstrate that the linear motor can compensate at an average RMSE of 0.002 mm for the texture on the printing surface. Latency in the system and algorithm are continuously being revised and used to optimize the mechanism. Currently, the algorithm can compensate adequately up to a velocity of 75 mm/s. These results show promise for this algorithm as an adaptive printing distance algorithm for real-time in-situ 3D bioprinting.
This project was supported in part by the NSF REU Site (Award #1950281) in the Department of Biomedical Engineering at Wake Forest University School of Medicine and by the Intuitive Technology Research Grant.
[1] Thai MT., et al., Robotic System for In Situ 3D Bioprinting. 2023, vol. 10(12)