A sensing system

Hello! I have posted an idea of mine which I think needs further explanation. First of all, are two sensors really necessary? At first sight, each sensor should be capable of measuring the liquid level independently with certain accuracy and precision. Thus, the tube gauge reads the liquid level directly whilst the flow rate gauge measures the liquid consumption. Then, the flow rate data must be multiplied by time and subsequently subtracted from the initial liquid quantity. Separately, each sensor suffers some disadvantages. For instance, if the fuel tank happens to lean to one side, the tube gauge will no longer be able to read the liquid level correctly. Alternatively, if a liquid leak takes place, the flow rate gauge won’t sense what is going on. Consequently, both sensors are complementary. It crosses my mind that both sensors data might be fused into a linear Kalman filter. In this way, one should be able to estimate the actual liquid level regardless disturbances of any kind. For instance, the raw data might be distorted by ripples, vibrations, and so on. The state vector contains the fluid volume and the control input vector contains the flow rate. The tube gauge capacitance might be converted into a sequence of rectangular pulses with different width by means of 555 timer working in monostable multivibrator mode. The flow rate gauge in turn also outputs rectangular pulses with fixed width. Therefore, reading the tube gauge data comes down to counting pulse width. Reading the flow rate gauge data boils down to counting pulses altogether. All these might be implemented by many microcontrollers. Last but not the least, a calibration procedure must be carried out so as to work out a relationship between each sensor raw data and the actual liquid level. Least squares technique is suitable to approximate a standard of liquid levels. To sum up, I think that not only does the sensing system require knowledge in mechanical design but also skills in control systems a