In-sensor shock intensity estimation
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.
1 . A sensor for detecting a shock event, the sensor comprising:
a machine learning core configured to continuously calculate an acceleration norm value based on a change in velocity along one or more axes from an accelerometer; and
a finite state machine circuit configured to:
compare the continuously calculated acceleration norm value to a first threshold value and, in response to the continuously calculated acceleration norm value being greater than the first threshold value, generate a first interrupt signal, and
compare the continuously calculated acceleration norm value to a second threshold value less than the first threshold value and, in response to the continuously calculated acceleration norm value being less than or equal to the second threshold value after the first interrupt signal is generated, generate a second interrupt signal,
wherein the machine learning core is further configured to calculate a maximum acceleration norm value between the first interrupt signal and the second interrupt signal and store the maximum acceleration norm value in a register,
wherein an application processor of a host device reads the maximum acceleration norm value after receiving the second interrupt signal from the sensor, and
wherein the maximum acceleration norm value is reset after a delay after the second interrupt signal is generated.
2 . The sensor of claim 1 , wherein the finite state machine circuit is further configured to calculate a shock event duration corresponding to the generating of the first interrupt signal and the second interrupt signal and store the shock event duration in the register.
3 . The sensor of claim 2 , wherein the application processor reads the maximum acceleration norm value and the shock event duration from the sensor.
4 . The sensor of claim 1 , wherein the acceleration norm value is calculated using a Euclidean norm, and wherein the maximum acceleration norm value is calculated using a recursive function.
5 . The sensor of claim 1 , further comprising a First-In-First-Out (FIFO) circuit, wherein the maximum acceleration norm value is copied into a memory element of the FIFO circuit in response to the second interrupt signal being generated, and wherein the application processor accesses the memory element of the FIFO circuit to read the maximum acceleration norm value.
6 . The sensor of claim 5 , wherein the delay is a first delay, wherein the finite state machine circuit is further configured to generate a third interrupt signal after a second delay less than the first delay, the third interrupt signal indicating to the application processor that the maximum acceleration norm value is copied into the memory element of the FIFO circuit.
7 . The sensor of claim 6 , wherein the finite state machine circuit is further configured to:
generate the first delay after the second interrupt signal is generated; and
generate a fourth interrupt signal after the first delay to indicate that the maximum acceleration norm value is reset and the sensor is monitoring for a next shock event.
8 . A system, comprising:
a sensor, comprising:
a machine learning core configured to continuously calculate an acceleration norm value based on a change in velocity along one or more axes from an accelerometer, and
a finite state machine circuit configured to:
compare the continuously calculated acceleration norm value to a first threshold value and, in response to the continuously calculated acceleration norm value being greater than the first threshold value, generate a first interrupt signal, and
compare the continuously calculated acceleration norm value to a second threshold value less than the first threshold value and, in response to the continuously calculated acceleration norm value being less than or equal to the second threshold value after the first interrupt signal is generated, generate a second interrupt signal,
wherein the machine learning core is further configured to calculate a maximum acceleration norm value between the first interrupt signal and the second interrupt signal and store the maximum acceleration norm value in a register, and
wherein the maximum acceleration norm value is reset after a delay after the second interrupt signal is generated; and
an application processor coupled to the sensor, the application processor configured to read the maximum acceleration norm value after receiving the second interrupt signal from the sensor.
9 . The system of claim 8 , wherein the sensor operates autonomously from the application processor.
10 . The system of claim 8 , wherein the application processor is configured to calculate a shock event duration corresponding to a duration between the receiving of the first interrupt signal and the second interrupt signal.
11 . The system of claim 8 , wherein the acceleration norm value (n) is calculated using the equation: n=√{square root over (αx 2 +βy 2 +γz 2 )}, wherein x, y, and z are different components of an acceleration vector, wherein α corresponds to a weight factor for the x direction, β corresponds to a weight factor for the y direction, and γ corresponds to a weight factor for the z direction, and wherein the weight factors are configurable.
12 . The system of claim 8 , wherein the sensor further comprises a First-In-First-Out (FIFO) circuit, wherein the maximum acceleration norm value is copied into a memory element of the FIFO circuit in response to the second interrupt signal being generated, and wherein the application processor is configured to access the memory element of the FIFO circuit to read the maximum acceleration norm value.
13 . The system of claim 12 , wherein the delay is a first delay, wherein the finite state machine circuit is further configured to generate a third interrupt signal after a second delay less than the first delay, the third interrupt signal indicating to the application processor that the maximum acceleration norm value is copied into the memory element of the FIFO circuit.
14 . The system of claim 13 , wherein the finite state machine circuit is further configured to:
generate the first delay after the second interrupt signal is generated; and
generate a fourth interrupt signal after the first delay to indicate that the maximum acceleration norm value is reset and the sensor is monitoring for a next shock event.
15 . A method for detecting a shock event, the method comprising:
continuously measuring, by a sensor in a host device, acceleration values along multiple directions;
continuously calculating, by the sensor, an acceleration norm value based on the acceleration values;
comparing, by the sensor, the continuously calculated acceleration norm value to a first threshold value and, in response to the continuously calculated acceleration norm value being greater than the first threshold value, generating a first interrupt signal;
comparing, by the sensor, the continuously calculated acceleration norm value to a second threshold value less than the first threshold value and, in response to the continuously calculated acceleration norm value being less than or equal to the second threshold value after the first interrupt signal is generated, generating a second interrupt signal;
calculating, by the sensor, a maximum acceleration norm value between the first interrupt signal and the second interrupt signal;
storing the maximum acceleration norm value in a register of the sensor, wherein an application processor of the host device reads the maximum acceleration norm value after receiving the second interrupt signal from the sensor; and
resetting the maximum acceleration norm value after a delay after the second interrupt signal is generated.
16 . The method of claim 15 , further comprising:
calculating, by the sensor, a shock event duration corresponding to the generating of the first interrupt signal and the second interrupt signal; and
storing the shock event duration in the register, wherein the application processor reads the shock event duration and the maximum acceleration norm value from the sensor.
17 . The method of claim 15 , wherein the acceleration norm value is calculated using a Euclidean norm, and wherein the maximum acceleration norm value is calculated using a recursive function.
18 . The method of claim 15 , further comprising copying the maximum acceleration norm value into a memory element of a FIFO circuit in the sensor in response to the second interrupt signal being generated, and wherein the application processor accesses the memory element of the FIFO circuit to read the maximum acceleration norm value.
19 . The method of claim 18 , wherein the delay is a first delay, the method further comprising generating a third interrupt signal after a second delay less than the first delay, the third interrupt signal indicating to the application processor that the maximum acceleration norm value is copied into the memory element of the FIFO circuit.
20 . The method of claim 19 , further comprising:
generating the first delay after the second interrupt signal is generated; and
generating a fourth interrupt signal after the first delay to indicate to the application processor that the maximum acceleration norm value is reset and the sensor is monitoring for a next shock event.