Supplier eBooks

13 ST/Industrial Sensing Solutions IIS3DHHC Three-Axis Linear Accelerometer • Ultra-low noise performance: 45μg/√Hz with Excellent stability over temperature (<0.4mg/°C) and time • 16-bit data output with SPI 4-wire digital output interface • Included in the 10-year longevity program LEARN MOREu STMicroelectronics Industrial Sensors eBook - Videos Video Title URL Predictive Maintenance & Monitoring Solution https://www.mouser.com/video/?videoId=6024440845001 IIS328DQ 3-Axis Linear Accelerometer https://www.mouser.com/video/?videoId=5088943196001 L20G20IS Two-Axis MEMS Gyroscope https://www.mouser.com/video/?videoId=5747109098001 Getting Started with LSM6DSOX https://www.mouser.com/video/?videoId=6097814870001 u IIS328DQ 3-Axis Linear Accelerometer application codes that need to get devel- oped, easing the implementation that helps reduce the development cycle, lowering power consumption, and reducing the total computing time. The following sections will provide an overview of significant embedded features. Finite State Machine & Machine Learning Core Until recently, there have been primarily two methods to implement and run sensor data algorithms. The first method is about embed- ding the solutions within the sensor. The main advantage of this method is the low power consumption. Another advantage is that this method requires only two dice in the same package (Sensing structure + dedicated elec- tronics IC). However, the number of embed- ded algorithms in the sensor is limited, and there is a low-level of flexibility when it comes to the configuration of these algorithms. The second method is the standard solution where data from sensors are collected, and the algorithms run on a low-power microcontroller (embedded in the same pack- age or external). In this case, all the data processing takes place in the microcontroller. Multiple algorithms can run in the microcontroller, therefore, offering a higher degree of flexibility. However, the main disadvantages are higher cost and higher power consumption compared to the first method described above. The third method of embedded solution takes a different and very innovative approach. This solution incorporates two configurable embedded modules that combine the advantages of both solutions discussed above. This third solution allows swift and effective implementation of mo- tion detection processing. It reduces the implementation efforts and consumes meager levels of power—20 to 100 times current consumption reduction respect to standard solutions. These two modules will get discussed further in the following sections. Finite State Machines The latest inertial sensors offered by ST offers a powerful configuration option to generate interrupt signals activat- ed by user-defined motion patterns. One can define the motion patterns by motion detection by programming the finite state machines embedded in the sensor. There Figure 6: Basic structure of programming of finite State Machines embedded in ST sensors. are 16 embedded state machines in the sensor to provide a high degree of flexibility to the user. The programming of the embedded FSMs takes place independently from each other for motion detection such as glance gesture, custom gestures, absolute wrist tilt, shake, double-shake detection, flip-up/down, device to ear, etc. These highly flexible embedded FSMs make the latest sensors a perfect solution for the implementation of various gesture recogni- tion algorithms at a negligible current consumption com- pared to running these algorithms on a microcontroller. At a very high level, one can define the implementation of the embedded state machines for various applications in three steps (Figure 6). The first step is the definition of the inputs to the state machines. That means to decide the sensor data that the FSMs will be using. The second step is the definition of states using commands, conditions, and pa- rameters. The third step is the definition of output of the FSMs. Figure 5: Fully molded package of MEMS pressure sensor with a cap wafer

• Cover