Embedded AI and MEMS Sensors Shape the Future

Bosch's MEMS sensor development started from automotive, and slowly expanded to the consumer field, which includes cell phones, wearable products, and then penetrated into the current IoT field.

Now, edge AI algorithms can also be integrated into small sensor products, because the processor power consumption has been very low, the processing power can also realize the operation of AI algorithms. So in the whole process, the sensor, a single piece of hardware, is slowly carrying more tasks to do smarter application scenarios.

What is edge AI?

The traditional AI carries the signal transmission, which passes the physical signal to the network side through wireless or other means, and after the network side receives the signals from hundreds of thousands or even tens of thousands of sensors, it can do modeling on big data and provide some intelligent responses and intelligent strategies in the cloud.

Benefits of AI Algorithms Integrated into Devices

First, customization or personalization can be done on the device side. The needs are different, and adapting to each user allows the sensor recognition accuracy to be adapted to each individual user.

Second, the sensor data does not need to be uploaded to the cloud, but only exists in the device itself, or even inside the sensor. This ensures the security of the sensor data, or the security of the user's data.

Third, real-time response is faster. There is no need to upload data to the cloud and then feed it back to the device side, responding directly in the device.

Fourth, further reduce the power consumption of the device to extend the battery life or the daily use of the battery.

BME 688 4-in-1 Sensor

As Bosch's latest 4-in-1 environmental sensor, BME 688 is also the world's smallest 4-in-1 sensor, which integrates temperature, air pressure, humidity and gas. Based on this environment sensor, it can provide users with a micro-environmental monitoring, and even do some micro-environmental weather forecast prediction, as well as with the smart home linkage and so on.

How does this sensor realize AI function?

The sensor is not integrated further with a controller or processor, so Bosch provides an AI learning development tool through the PC side, and users can do some customization based on this tool, or application scene customization.

According to the case study, because the BME688 can detect different gases, mainly VOC gases, how to do gas differentiation? The first step is that the user defines his gas type, for example, if he wants to detect the odor of a certain kind of coffee, he can base on the target gas of coffee and let the sensor to collect the signal of the target gas. After the data is collected, there is a corresponding simplified tool on the PC side that allows the user to do data characterization based on the collected sensor signal of the target gas. After capturing the data features, the tool is used to generate a data model to identify the target gas. Finally, the generated model is integrated into the customer's algorithm, which can be combined with the physical signals of the sensor to perform functional detection in real application scenarios.

VOC gases are used in a wide range of applications, such as detection of forest wildfires, detection of harmful gases, detection of baby diapers, detection of food freshness in refrigerators, and even detection of outdoor air indices.

The BME688 works by heating the metal oxide inside the sensor to react with the target gas, which has different characteristics at different temperatures. So by collecting multiple temperature points, corresponding to the resistance value of the reduction reaction, the characteristics can be captured, and can be differentiated after capturing.

For example, if there are three volatile gases, the features captured are not the same, so we can quickly model the features through the resistance values reported by the sensors at different temperature points.

If you like this article, click into perceptive-ic.com to see more.