|
A Prediction to the Position of MLC with Machine Learning Algorithms |
|
Pilsu Kim, Manjin Ha, Byung Jun Min(Chubgbuk National University Hospital, Korea) |
| The positional information of MLC, Gantry, and jaw is important to make a sophisticated treatment. A few of ML algorithms were tested to predict the position of MLC. The input data related to the motion were better than the only position of MLC as the input data. The MAE was 0.14595 when the ML model is the multi variate linear regression. As the target data were the field of view, the best MAE was 1.49. The various input data are better than the only position input data. |
|
|
3D Real World Representation with Octomap and Euclidean Distance Field |
|
Jeongjun Choi(Seoul National University, Korea), Boseong Jeon(Seould National University, Korea), H. Jin Kim(Seoul National University, Korea) |
| UAV(Unmanned Aerial Vehicle) navigation has become increasingly important over the last decade. It could
be hazardous for UAV to fly in GPS-denied region, so it is important to build a map that ensures safety at the same time as
completing the mission. Octomap is such a tool to efficiently build a map, representing both free and unknown areas with
probability. In this work, we build an octomap with Euclidean Distance Field(EDF) in real 3D cluttered environment and
adjust some parameters to avoid an excessive conservative flight. |
|
|
Measurement of the Catheter Motion for Catheter Simulation |
|
Chengjie Li, Doo Yong Lee(KAIST, Korea) |
| This paper presents a method to measure the catheter motion for a simulation of medical catheterization procedures. An optical navigation sensor used in a common computer mouse is utilized to measure the translational and rotational velocity of the clinical catheter. Catheter passage is designed to keep the catheter inside the effective sensing area of the optical sensor. Experiment results show that the developed device can measure the velocity up to 150 mm/s in translation and 6 rad/s in rotation. |
|
|
FMCW Lidar Design using Biaxial Polarization Optics and Improvement of Intereference Frequency SNR |
|
Jinseo Hong(GIST, Korea), Kyihwan Park, Jubong Lee(Gwangju Institute of Science and Technology, Korea) |
| Lidar has the advantage of being able to obtain 2D and 3D images directly regardless of the presence of lighting. So it is being used in many fields such as robotics, autonomous vehicles and transporation infrastructure. However, TOF lidar, which currently accounts for most of the market, is vulnerable to environments such as snowy, rainy, foggy weather and direct sunlight. In this study, we design an FMCW type lidar that is strong in harsh environments by using a biaxial polarizaiton optical system. It eliminates internal reflection which induce great improvement of internal frequency SNR. |
|
|
DQN-based Automotive Cabin Temperature Control System |
|
Seohee Han, Changsun Ahn(Pusan National University, Korea) |
| The market for eco-friendly vehicle is growing due to the strengthening of greenhouse emission regulation. In the case of electric vehicle, optimization of energy efficiency used in integrated thermal management is important because battery is the only energy source. In this work, we propose Deep Q-Network based controller to improve the air conditioning energy efficiency of electric vehicle. DQN is not necessarily required for a control model and can control the system by state of environment. The DQN-based controller shows improvement of energy efficiency and reduction of temperature error. Also, it can be confirmed that it has robustness to driving cycle. |
|
|
Initial Investigation on the Friction Estimation of a Bowden-cable Measuring Electrical Resistance |
|
Useok Jeong(Korea Institute of Industrial Technology, Korea) |
| this paper focuses on the initial investigation on the bending angle estimation of Bowden-cable, which is closely related to cable friction, by measuring the electrical resistance of the actuation wire. The proposed method measures the change of electrical resistance of the Bowden-cable’s actuation wire along the routing path while the tension is applied to the cable system. We propose the bending angle estimation algorithm of the cable’s output port based on the friction distribution model of the Bowden-cable using the measured tension on the input port and the wire’s resistance. |
|