Main Work: 1. Instead of triangulation, use machine learning model to determine the threshold of RSSI which always unstable. 2. Upgrade the TF1 to TF2, and improve model accuracy in lots of cases like Mackay Memorial Hospital(馬偕), BMW, Leaspy(宇博), Hannstar(瀚宇彩晶) and etc. 3. Conduct experiment and write test for new product, for beacon transmitted by BLE. 3. System maintenance and repair Other research: 1. For the propose of improve locating accuracy, search the object detection 2. Demo Raspberry Pi 4 with Google Coral Accelerator to run related object detection models

Main Work: 1. AOI inspection of panels and light guide plates, and helps factories to reduce labor costs. 2. Use Energy-based model to achieve automatic detection finding the defect on panels. 3. Then, determining the threshold by OpenCV 4. More, Semi-Supervised learning can help to define what is the defect like the cognition by operators, and the concept of out-of-distribution may be the good way to determine the threshold. 5. With a small number of light guide plates, the results of the algorithm at the beginning of the cutting test: accuracy 0.814 (base: 0.643), F1 score: 0.691 6. Applying patent for the above algorithm Others: 1. Train GPU efficiently, and build docker file for personal project in order to separate others environment. 2. Build the automatic daily mail for reporting the data of production by the factory, and the condition of the hardware.