Q1. Briefly introduce your graduation project?
The title of my graduation project is "Design of a Paper Counter Based on Capacitive Sensors". In general, it can be roughly divided into the following steps. The first step is to consult relevant information, come up with several implementation plans, and analyze the feasibility and accuracy of the plan. The second step is to build the hardware structure, including mechanical and electrical control parts. Step 3: Carry out Kalman filter processing on the collected capacitance data, collect the raw data on the scene of the sensor, select the appropriate Kalman noise covariance Q and R with the help of matlab tools , and perform regression analysis on the filtered data. The least squares method is used to fit the data piecewise. And constantly debug and optimize.
The fourth step is to write a UI interface for better human-computer interaction.
Q2. Why did you choose this topic, what is the significance of the research?
There are two reasons:
first, I feel that this topic is of some value to society; the current paper counting methods mainly include manual counting and mechanical equipment counting, and there are also weights to reflect the amount of paper. These methods have certain disadvantages, such as inaccurate counting and wear on paper. The research on this subject can solve these problems well.
The second is that I have learned something about this subject. There were some problems in the previous design process, and the existing problems were solved through this design. At the same time, when doing this topic, I thought of using some new methods, such as using Kalman filter to deal with Gaussian noise, least square method to solve the curve regression problem, and better noise shielding. I think I can learn a lot in this design.
Significance : First of all, it is innovative to measure the quantity of paper by measuring capacitance. At present, there is basically no such product on the market. Secondly, measuring the quantity of paper by measuring capacitance can solve some problems in life that other methods cannot solve For example, to measure the amount of paper in the printer, some methods based on machinery or vision cannot solve this problem at all, and the problem can be solved well by reacting the paper with the capacitive electrode plate.
And this method is simple and highly feasible.
Q4. What problems did you encounter? How to solve it?
1. Noise interference problem, when the sensor is measuring, the original data will fluctuate, and the repeatability of multiple measurements is low.
2. Selection of data processing methods. .
Solution:
1. Considering the sensor noise problem, the Kalman filter is used to filter the data, which can filter out the noise well, and also improve the repeatability to a certain extent. At the same time, use another channel to measure the blank group to reduce the environment Impact on data accuracy.
2. Due to the limited floating-point computing power of the previous microcontroller, the STM32F4 series chip with FPU was replaced, and the floating-point capability was significantly improved. The fifth-order polynomial was used to process the data fitting and achieved very good results.
Q6. What are the expected results?
There is no error within 100 sheets after calibration, and the paper error within 100 to 200 does not exceed 2%.