Signal processing

Signal processing: analyzing and interpreting

As the number of digital devices has increased in the last few years, the use of signal processing has increased with it. Currently, it is an essential domain when you are working with new technologies. This science helps to analyze and interpret information coming from sensors in real-time.

Like a human operator who interacts with his environment depending on the kind of information received, “smart devices” do the same by processing a huge number of pieces of information (temperature, position, speed, pressure, flow, voice…). The measurements, which are obtained thanks to sensors and evolve continuously over time, are called “signals”.

Often mixed, noised or biased, signals require going through some processes before being used by an algorithm. Signal processing consists in smoothing values, filtering them, cleaning absurd values off and performing any action to enhance signal quality.

In addition to processing information coming from the sensors, we are able to combine it with our knowledge of the system behavior, so as to evaluate unmeasurable data: this is a “state observer” and it is one of the many possible approaches with signal processing.

Acsystème has designed and implemented the most efficient and effective ways to stick to the system specifications. From simple filtering to the most advanced statistic processing and including time-frequency analysis, Acsystème’s specialists set up these kinds of processes, calibrate settings, and judge the results generated from actual signals.

Here are some examples of signal processing techniques:

• discrete-time and continuous filtering, Kalman filters…
• time/frequency analysis, spectrum analysis…
• online or offline algorithm