Implementation of an adaptive asynchronous signal sampler, which changes the bias of the signal in real time. The project is based on a special sampling algorithm and improves the regular one by reducing the number of samples without losing information.
Today ADCs are mainly using periodic sampling and quantization of signals. An energy efficient type of Asynchronous ADC is a time encoding machine, which operates without relying on the global clock.
The focus in this project is on AIF-TEM sampler, which is Adaptive Integrate and fire time encoding machine.
The AIF-TEM is based on Integrator which sums the input signal, and comparator that compares the integrator output and a threshold constant – 𝛿.
Every time the integrator output reaches the threshold value of 𝛿, a trigger signal is sent, and the integrator resets. The sampling itself is done by measuring the time instance between two following triggers.
To keep the input of the integrator positive, a bias is added to the signal, as it is required for the proper functioning of the IF-TEM. In our project, the bias changes in time and is estimated iteratively using a special algorithm based on the signal and on previous bias values.
The main goal in the project is to write compatible code for this algorithm, load it on hardware containing the integrator and comparator, and to test the bias estimation during the system’s operation.
