| ObjectivesAs one of the main ways to restore missing teeth,dental implants’clinical success is related to the final osseointegration.During the preparation of the implant sites,the heat generated by the drilling needle cutting the bone tissue may lead to high local temperature,which may cause bone damage and affect the osseointegration and eventually lead to implant failure.In traditional freehand dental implant surgery,the surgeon cannot directly obtain the temperature inside the cavity,but can only infer the heat generation by feeling the resistance through drills,and then make appropriate adjustments based on experience to avoid excessive heat on the bone tissue.This process totally depends on the operators’subjective factors and does not meet the needs of the different bone qualities in the operative area.Relying on its force servo function,the dental implant robot could precisely adjust the surgical parameters according to the real-time acquired instrument resistance,reducing the tissue damage and implant failure caused by the surgeons’subjective factors and even improving the intelligence of the dental implant robot.In this study,we intend to construct an in vitro research model based on bovine ribs,find a suitable method for measuring and evaluating the temperature and force values at the end of the implant handpiece,and then explore the relationship between the three-dimensional directional feedback force at the end of the drill pin and the heat production during the implant surgery of the dental implant robot,and try to establish a prediction model for the three-dimensional directional feedback force to predict the intraoperative temperature rise,eventually lay the foundation for the autonomous temperature control of the dental implant robot during the surgery.Material and methods1.Establish an experimental model suitable for the dental implant robot platform.Using a modified bench clamp device,bovine ribs,combined with the dental implant robot system to establish an in vitro implant research platform,and to test the precision of preparation with the device.2.Based on the in vitro study model,the acquisition and evaluation of temperature and force data were determined.Experiments were designed to compare thermocouples and infrared thermography and establish a suitable temperature and force evaluation method for this study.3.By introducing several parameters that affect heat production in clinical practice(cortical bone thickness,drill needle diameter,drill needle speed,and drill needle wear)to setting up different preparation scenarios,so that multiple sets of temperature and force data were obtained and analyzed for correlation.Firstly,the correlation between temperature and force values was tested using Pearson correlation.And second,a multiple linear regression model was established for further analysis.Finally a temperature prediction model was established by 4 machining learning(ML)algorithms(support vector regression[SVR],ridge regression[RR],extreme gradient boosting[XGboost]and artificial neural network[ANN]).Results1.3D design,CNC cutting,and 3D printing technologies were used to transform the bench clamp device into a new one that can be recognized and tracked in real time by the vision system of the dental implant robot.The accuracy of the device for preparation in the dental implant robotic platform was tested experimentally.And the results showed that the average value of the thinnest thickness of the hole sidewall was 0.51±0.08 mm,which was not statistically different from the set value of 0.5 mm and facilitated the accurate temperature measurement by the temperature measurement device infrared thermography and thermocouple.Using this device,no coolant flow to the temperature measurement surface of the bone block was observed in the experiment.2.The results of temperature measurement by thermocouple and infrared thermography showed that the infrared thermography responded faster to temperature changes and measured higher peaks,and the average temperature measured was about12.2°C higher than that of the thermocouple.The force values monitored by force sensor were transformed into force data consistent with the implantation drill needle axis using the spatial vector transformation equation.The force data(Fx,Fy,Fz,Tx,Ty,Tz)were transformed into new variables combined force and time(Fx-t,Fy-t,Fz-t,Tx-t,Ty-t,Tz-t)using the method of calculating the area under the curve of the force sensor data.3.It was observed that the changes in force values were correlated with the changes in temperature for different clinical parameters(cortical bone thickness,drill needle diameter,drill needle speed,and drill needle wear).The results of Pearson correlation analysis showed that the six force variables(Fx-t,Fy-t,Fz-t,Tx-t,Ty-t,Tz-t)were strongly correlated with the maximum temperature difference(?T)with R~2≥0.8.After excluding the multicollinearity,a multiple linear regression model was developed.The results of the analysis showed that the multiple linear regression model incorporating the three force variables(Fz-t,Ty-t,Tz-t)had a better performance with R~2of 0.837.Among the four machine learning models(RR,SVR,XGboost and ANN),the SVR had the best performance,because of its smaller MAE and MAPE.Conclusion1.In this study,an in vitro bovine rib bone experimental model was successfully established by combining the high-precision infrared vision system of the autonomous dental implant robot and using digital technology to modify the bench clamp device.The model can be accurately positioned by the vision system of the dental implant robot through the registration point and the assembled Marker with good preperation accuracy.The model was able to solve the problem of coolant affecting the temperature measurement.In the same experimental group,after the model was calibrated and aligned in one operation,the experiment can be performed continuously by constantly replacing the bone block,which is conducive to the large sample experiment.2.By comparing the two commonly used temperature measurement devices,it was found that the infrared thermography has faster response and more accurate measurement results,which is more suitable for this study.Therefore,the infrared thermography was used as a temperature measurement tool in the subsequent study.The maximum temperature difference(△T)was determined as the temperature evaluation index.The force value data was transformed into a new variable and used as the force value evaluation index for this study.3.Force value data and temperature data had similar trends under different clinical parameter variations.Pearson correlation analysis proved that the force value variable was indeed correlated with the maximum temperature rise.Although the multiple linear regression model showed better predictive performance,four machine learning models were further analyzed to better present the relationship.Finally,a prediction model using force sensor monitoring values to predict the temperature rise in preparation sites was successfully constructed under the existing experimental conditions,which laid a good foundation for further realization of autonomous control of heat production by the dental implant robot during implant surgery. |
PDF Full Text Request