In this project we have an MPU6050 accelerometer communicating with our MCU over I2C. Whenever A noise is detected above a programmable threshold, an interrupt is raised and the mcu starts sending accelerometer data.

The calibration takes a predefined number of samples (1024) and computes an average for each of the axis (denoted as E[x], E[y], E[z]). In order to make sure the samples are good, we make sure the variance of the samples is below a certain threshold. After sampling, we take the vector (E[x], E[y], E[z]) and consider it our true Z vector. In order to compute our true X & Y vectors we need to compute 2 orthogonal vectors to Z. The X vector can be achieved easily via taking the vector (E[y], -E[x], 0).

The Y vector can be computed as the cross product of Z & X, which is (E[x] * E[z], E[y] * E[y], -(E[x]^2 + E[y]^2)).

Now all that is left to do is create a basis change matrix and multiply any reading we have with this matrix.

On register 0x1F of MPU6050, we can define a threshold for the MPU's interrupt pin to turn high, subsequently causing an interrupt in the MCU. once the interrupt is received, the MCU takes samples in a predefined interval, processes them, and batches them up using collector.h. Once a batch is full, it is sent via mqtt to a broker.

Once the samples variance falls below a threshold, the event is over, the MPU's interrupt is cleared and the MCU goes idle.

The MCU is subscribed to a certain topic (tonto2/set), and upon receiving a message will update MPU6050's threshold register (0x1F) to the specified value. Thus, the sensitivity can be remotely adjusted.

Every 30 seconds of inactivity, a ping is sent to the server in order to keep the connection alive. This is implemented using a timer & an interrupt.

Pressing BTN0 will force a calibration event. Pressing BTN1 will make the cat mewo.