With over 30 years’ of experience in developing products for factory automation and industrial applications, STMicroelectronics is a key player in the Smart Industry space. During IIOT India 2020 held in Gurgaon earlier this month, ELE Times caught up with Vishal Goyal, Group Manager, Technical Marketing, South Asia and India,STMicroelectronics’ and Vinay Thapliyal, Technical Marketing Manager, MCD, India, STMicroelectronics and learnt about company’s latest products and solutions in the field of Industrial IoT showcased there. The technology showcase was focused on demos from Industry 4.0, IoT for smart homes and cities and IoT for smart things including: AI Powered Object Classification demo; AI Character Recognition demo; STM32WL Long range Demo; STM32L5 devices; BLE Mesh; Sigfox Monarch Reference Design and BlueNRG Tile., Vishal Goyal elaborated on key Automation demos as follows
SensorTile.box: SensorTile.box is a ready-to-use box kit with wireless IoT and wearable sensor platform to help use and develop apps based on remote motion and environmental sensor data, regardless of your level of expertise. The SensorTile.box board fits into a small
plastic box with a long-life rechargeable battery, and the ST BLE Sensor app on your smartphone connects via Bluetooth to the board and allows you to immediately begin using the wide range of default IoT and wearable sensor applications. In Expert Mode, you can build customs apps from your selection of SensorTile.box sensors, operating parameters, data and output types, and special functions and algorithms available.
BlueNRG Tile: BlueNRG-Tile or STEVAL-BCN002V1D Bluetooth LE enabled sensor node development kit features the multi-sensor board based on BlueNRG-2 SoC Bluetooth Low Energy application processor. This sensor boardhas accelerometer, gyroscope, magnetometer, pressure, humidity, Time-of-Flight and microphone sensors, and is powered by a common CR2032 coin battery. The sensor board communicates with a Bluetooth LE enabled smartphone running the ST BLE app, available on Google Play and iTunes stores.
BlueNRG Mesh: BlueNRG-Mesh is a software solution for connecting multiple BLE (Bluetooth low energy) devices in Mesh networks for Internet of Things (IoT) solutions. It enables true two-way communication between Bluetooth-enabled devices in powerful, secure, integrated and range-extending Mesh networks. The demo included Mesh connected Smart RGB bulbs, BlueRNG-Tile & SensorTile.box for sensor data and Mobile to provision, control and display the network. BlueNRG-Mesh is a software solution for connecting multiple BLE (Bluetooth low energy) devices in Mesh networks for Internet of Things (IoT) solutions. It enables true two-way communication between Bluetooth-enabled devices in powerful, secure, integrated and range-extending Mesh networks
Based on the conversation following is the informal narration of demo:
Vishal Goyal: There are three pillars of ST’s industrial strategy. They are Automation, Power & Energy and Motor Control. In this booth our main focus is around Automation. For Automation we are showcasing our ability to integrate the sensors, process the sensor data and send them via connectivity for further analysis. The key technology around automation is Bluetooth Low Energy. We are showing Mesh over BLE with which you can automate your office, factory and home. You can connect several device to form a Mesh network in your premises. The Mesh network is two-way communication so any node can talk to any other
node. The Mesh support by ST is fully interoperable and Standard SIG Mesh solution. In the displayed solution three RGB Bulbs are connected in Mesh network which is controlled using
a Mobile tablet. We are also displaying sensors data using Mesh connected Kits – BlueNRG Tile and sensrotile.box. The advantage of Mesh is that in increases the range of BLE so you need not to be in range of individual devices yet control and sense them. We have enabled a ecosystem through we can work with any LED fixtures manufacturers and help them convert into Smart LED fixtures.
ELE Times: How do you control it? Is it possible through an application or any device?
Vishal Goyal: It can be controlled through tablet as I am controlling it. All these lights can be controlled at one go or can be controlled individually as well. We have integrated some sensors in this network. I can control the lights with the sensors. These sensors are measuring the vibration through which it can be turned on or off.
ELE Times: And if there is no vibration, what would happen in that case?
Vishal Goyal: You define what you want to do. We have integrated this particular functionality, but you can control them manually as well. These sensors can be put on a window. And if somebody tries to break into a window all the lights will turn on and the alarm will also turn on. If a person is moving, according to his location using proximity sensor, that particular lights will turn on and off. Similarly you can control the Lights based on other sensor inputs such as humidity and temperature.
ELE Times: If we want to control it remotely then do we need to download any app?
Vishal Goyal: If you are in the same premises you need one app to control it or you need to have some sort of switches like shown here. If you want to control them from outside the premises then you need to have one gateway which can take this data to the cloud and you can access the cloud through other means such as tablet or mobile. This is all about technology part. Enabling for the adoption of the technology: We provide several evaluation boards which customers can use for testing. Evaluation boards have integration of sensors inside. Apart from evaluation boards we also provide ready to use firmware package, Android and iOS SDKs. So one can get a mature starting point to work on our solution. We also have ecosystem partners who can provide more ready to manufacture solution. If one is not expert in RF technology, they can get the modules from ST or our partner companies. Modules can be integrated easily as all the RF related components are already mounted and complexity already address by us. We are providing several example applications for easy demonstration and integration.
ELE Times: Is the technology already in the market or is this something newly introduced?
Vishal Goyal: This technology is already there in the market but what we are introducing is a standard mesh solution. Most of the solution in the market today are proprietary or custom Mesh. So a switch from one particular company will not work with Bulb from a different company. This way customers will get locked-in with one particular manufacturer or ecosystem. What we are offering is standard SIG solution so you can buy the products from any company which can have chip set from any maker, even from our competitors. If they are using a standard solution, they will all talk to each other. Like Wi-Fi. This is what we are offering. This is something new and has not evolved so far.
Vinay Thapliyal, Technical Marketing Manager, MCD, India, STMicroelectronics shared on Security & AI demos AI Powered Object Classification demo: Image classification using the STM32H7 discovery kit with STM32F4DIS-CAM camera module. Trained for 18 classes food recognition. AI Character Recognition demo: New ultra-low power STM32L4/L5 discovery board. Human Machine Interaction (HMI) demo using the touchscreen for AI-enabled hand-writing character recognition.
Vinay Thapliyal: STMicro is showcasing a product line in security and Artificial Intelligence (AI). We are working on the ‘AI at the edge”. The idea of the edge is that – most of the time when we talk about AI, generally we talk about the cloud. But in forthcoming scenarios there are going to be a lot of cases when we are talking about the nodes that need some connectivity. If there is no good connectivity between the node and the cloud, in that case, you need some good time intelligence on those nodes. Earlier we used to call them smart?All the Nodes are doing well.
Now we’re talking about them getting more intelligence. We are putting AI inside them. Though, this will have a basic AI, this cannot process a big data. But we’re talking about some sort of intelligence which can go on to the individual node. Imagine if there are hundreds of case of nodes connecting to your cloud, and if they’re sensing the data which is useful for the cloud then the need of the bandwidth will be lesser compared to what it is today. For example if I have to recognise one food item and if this node is a dumb device and does not have more intelligence, then it has to capture the image, send it to the cloud and ask cloud to recognise the object. Cloud has the intelligence and can recognise the food item.
But if the node is a bit more intelligent and is not connected to any cloud, we call them ‘AI at the edge’ objects. In this demo, we have trained this node to recognise the food. It has been trained for eighteen different type of food and is able to capture the image and recognise on its own. It only sends the relevant data to the cloud.
ELE Times: In case if there is a weird kind of image, how will it recognise?
Vinay Thapliyal: As I mentioned earlier, it is AI at the edge where we have very limited intelligence. But if there is some image, which it is not trained at this moment those images can be sent back to the cloud for its recognition. Through some mechanism, cloud can retrain the model on what to do if this kind of image comes in the future. This could be a closed loop system that may not work in real time. For this ST has developed one tool which is called STM32Cube.AI. This is for developers. Developer can train the AI models using any of available models. There are a lot of available models in the market. These models create a big image that cannot fit in a small microcontroller. Our tool i.e. STM32cube.AI helps to optimize this model and fit inside a small microcontroller. In this case, food recognition, we have trained this model that could fit inside a small footprint of let’s say around 200 kilobyte of flash so that this can easily fit in the microcontrollers. And then it is able to do real time recognition. This can be called a real time AI at the edge. In the case of wearable – If I am walking, running, I am on car or outside car or in the park – you need a small AI. You don’t need bigger AI. So depending on the use case, you have to define your own thing straight. Similarly, in the case of character recognition you can write a character and it gets recognised.
ELE Times: So any number of characters can be recognised?
Vinay Thapliyal: Yes, depending on application it can be trained to recognize required characters. In this demo, this is just the idea that you are writing something and you can take the action or do something on this. This is just an example of the character recognition and AI at the edge. Other things which we are showcasing is the Connectivity. In connectivity we have launched world’s first SoC STM32WL which is based on Long Range technology i.e. LoRa which is embedded with the microcontroller i.e. STM32L4. Along with this, we have embedded the transceiver which is supporting the LoRa, and other modulation techniques and the security needed for good and secured communication in long distances. So this is a technology can be used in Industrial automation, street lighting, gas metering, water metering, asset tracking and many more applications. This technology has very low power consumption so can work with small battery. With the help of a small battery this kind of connectivity take the signal to longer distance. You can have a small gateway at peak premises and can take it back to cloud. That way this technology is emerging and ST has launched its new product. On the security side ST is working quite strongly and recently we have launched STM32L5. This is the device, which is highly secured for the edge computing and edge type of processing.
Vinay Thapliyal Vishal Goyal