Methods and systems are disclosed for sensor-based classification. A collection of subset models that assign one of a plurality of class labels to input data is stored so that a first subset model of the model collection having at least two of the plurality of class labels is loaded in a sensor processing unit based at least in part on a first context. A first set of data is obtained from a sensor of the sensor processing unit and a class label is output for the first set of data from the first subset model. Then at least one subsequent subset model of the model collection is loaded in the sensor processing unit based at least in part on at least one subsequent context. Correspondingly, at least one subsequent set of data is obtained from the sensor so that a class label from the at least one subsequent subset model is output for the at least one subsequent set of data.

Disclosed is a MEMS device, comprising: a movable electrode plate; a first electrode plate and a first feedback electrode plate located on a first side of the movable electrode plate; a second electrode plate and a second feedback electrode plate located on a second side of the movable electrode plate. The first electrode plate, the first feedback electrode plate, the second electrode plate, the second feedback electrode plate respectively form a first capacitor, a first feedback capacitor, a second capacitor and a second feedback capacitor with the movable electrode plate. The first and the second capacitors are coupled to a detection circuit for performing differential detection on the first and the second capacitors; the first feedback capacitor and the second feedback capacitor are coupled to a feedback circuit for eliminating nonlinear relationship between an output voltage of the detection circuit and a displacement of the movable electrode plate.

According to an embodiment, a sensor including a machine learning core (MLC) and a finite state machine (FSM) circuit for detecting a shock event is provided. The MLC continuously calculates a value based on the change in velocity. The FSM circuit compares the value to a first threshold and generates a first interrupt if it is greater than the first threshold. The FSM circuit then compares the value to a second threshold less than the first threshold and generates a second interrupt if it is less than or equal to the second threshold after the first interrupt. The MLC calculates a maximum value between the first and second interrupts and stores it in a register, which is read by an application processor of a host device after receiving the second interrupt. The maximum acceleration norm value is reset after a delay after the second interrupt is generated.