AuthorLi Ying: (WeChat ID: ilovekm2008) Senior Editor of Smart Product Circle
What is most important in the Internet of Things? Data, data, and more data. Data will create value and change business models and competitiveness. But where does the data come from? The only answer is sensors. Building the Internet of Things requires sensors to collect data, which is then transmitted to the cloud via wireless technology. In terms of the perception layer, transmission layer, and application layer of the Internet of Things, the transmission layer has essentially been “cloudified”. The integration of cloud and communication modules has reached a high level of uniformity, whether approached from Bluetooth modules, Wi-Fi modules, or Zigbee modules, PLCs, etc., all have been tested countless times. As the core of the “end” in the cloud, pipe, and end chain, sensors still have much potential, and it may be time to consider how to leverage this “momentum” to soar into the “cloud”.
Leveraging the “Cloud” to Go Global
As the “source” of data, accurate data collection is crucial. Wang Yongtao, founder and CEO of cloud sensor manufacturer Maxense, told reporters, From a technical perspective, data accuracy = calibration + adjustment. In terms of calibration, currently, 90% of manufacturers in the market use static calibration methods, while Maxense employs dynamic calibration methods. In terms of adjustment, although local sensors can achieve self-calibration and self-adjustment to suppress drift, this is not enough; cloud-based calibration support is also needed.
Wang Yongtao gave an example: for instance, the PM2.5 levels in Beijing and Shanghai are both 300, but the proportions of PM10 and PM2.5 are different. A single sensor cannot analyze this; therefore, a cloud-based online calibration engine is needed to analyze the data, determine the environment and scenario of the sensor, generate a data report, and derive a calibration coefficient, which is then sent back to the sensor for calibration. Additionally, data from a single sensor can be biased; if 100 sensors are deployed in an area, using statistical methods to collect data from multiple sensors can provide mutual reference, and more data can calibrate the data from a single sensor. Cloud-based calibration will play a role in regional and scenario-based mutual reference.
Currently, Maxense has launched nine types of sensors, including temperature, atmospheric pressure, noise, illuminance, VOCs, humidity, formaldehyde, carbon dioxide, and PM2.5 sensors. Wang Yongtao explained that the characteristics of cloud sensors need to be customized based on applications. Maxense modularizes the sensors by writing optimized algorithms into a proprietary 16-bit MCU and integrating wireless connection chips to transmit data to the cloud. These sensors can be freely combined in a modular fashion to create five-in-one or even six-in-one configurations.
Once the data is available, how to effectively utilize it to deliver value to users is essential, and the support and services of the cloud platform are indispensable. Chen Peng, co-founder of AbleCloud, which provides the cloud platform for Maxense, stated, AbleCloud has built a rapid development framework for IoT cloud services starting from PaaS, allowing manufacturers’ existing R&D teams to easily and quickly develop the cloud functions required for hardware, and then develop SaaS services on top of that. The SaaS layer provided by AbleCloud offers a variety of service components, such as account system services, device management services, device sharing services, OTA management, event notification services, scheduled task engines, device sharing, and a series of other services. Therefore, the overall advantage lies in distributed scalability and modularity, allowing manufacturers to customize and develop their own cloud services without reinventing the wheel from scratch.
Currently, Maxense’s cloud sensors have been successfully applied in fields such as home appliances, real estate, and industrial manufacturing, playing a role in hardware such as water purifiers and fresh air systems. We can help traditional manufacturers quickly transition from non-intelligent solutions to intelligent solutions, facilitating industrial upgrades. Starting from scratch would require hiring a team of about ten software and hardware engineers and spending eight months to complete the project, as sensor calibration, algorithms, and calibration are very complex. Maxense provides a complete IoT perception architecture, where the sensors are directly connected to the cloud, enabling data interaction, with corresponding interfaces and original code open, and providing users with an SDK. Data can be read in any scenario, whether on mobile or PC, simply by embedding the module into the hardware product, thus only requiring one or two engineers to develop intelligent hardware.” Wang Yongtao pointed out, “As a result, customers can significantly shorten the development process for intelligent hardware and reduce development costs, saving several months of development time and nearly 90% of costs.” Currently, Maxense has received orders for 1 million units.
“Insight” Value
When sensors take flight with the “cloud”, it brings data to life, which means they have taken to the skies. The future is big data in the Internet of Things, and the demand remains unchanged. Currently, many people do not understand the value of data; we aim to become the owners of data and then gain insights into its value. Based on this, whether transforming into services or monetizing, there are many models to pursue, but one must first possess the data.” Wang Yongtao emphasized, “Maxense has cases in various fields and has become a pioneer in IoT development platforms. Data collection expertise is our eternal positioning.”
Moreover, realizing scene linkage is where the value of intelligent hardware services lies, which requires bidirectional interaction between sensor data and cloud data. At the sensor level, Wang Yongtao stated that Maxense provides two methods: one is to associate the execution mechanism with the sensor, where certain data sources have direct connections and can retrieve data from the cloud API, which is open; the second is that cloud sensors can directly send data to the execution mechanism locally, using wired, wireless, or broadcast methods, all of which Maxense can provide and implement.
Chen Peng pointed out that AbleCloud provides a universal platform, whether on the cloud or firmware side, which is universal. Commands issued from the cloud to devices will be transmitted to the network module, which will then relay the commands to the MCU in the sensor module. After processing, the MCU will reply to the cloud, thus forming a closed-loop service.
The convenience and customizability of cloud service development are also key to driving manufacturers to quickly access the cloud and release data value. “In addition to basic universal cloud services, AbleCloud creatively adopts a microservices architecture to provide a cloud service development framework and operating environment, namely UDS (User Defined Service). Manufacturers can easily and quickly develop their own business-related cloud services based on our platform without worrying about the operation, upgrade, maintenance, and expansion of these custom services. In short, UDS provides manufacturers with the ability to develop custom cloud services while significantly reducing their operational costs, making it zero maintenance for manufacturers.” Chen Peng introduced, “Most customers need such backend services, such as calculating PM2.5 values for air detection, processing data for water quality and quantity in water purifiers, etc. These different customers have completely different business logic, but based on the framework we provide, backend service development can be completed in two to three days.”
The Road Ahead
Sensors in the era of the Internet of Things, cloud computing, and big data have changed dramatically. “If sensor data cannot exist in the cloud and become big data, its individual value will be very small. The transition of sensors to cloud sensors allows data to be stored and analyzed, which is where value is realized.” Wang Yongtao mentioned.
Moreover, the transformation of sensors themselves is underway. Wang Yongtao stated that, on one hand, low power consumption in sensors is a trend, which will require smaller sizes, but cannot be endlessly miniaturized. For example, the measurement volume of PM2.5 sensors is positively correlated with the results. For data collection, accuracy and stability are more important than size. The future development will depend on whether the market capacity and driving force are large enough to generate new momentum and disruptive technologies, such as electronic compasses that were over $20 ten years ago and were large, but now cost only $1 and can be measured in millimeters, because they have reached a shipment volume in the hundreds of millions, prompting manufacturers to invest heavily. On the other hand, multidimensional measurement of sensors and sensor matrices are also noteworthy development directions.
“Sensors involve electrochemistry, optics, and semiconductor materials, requiring interdisciplinary integration. The technical means used, the exploration paths, how to modularize, how to combine applications to provide complete solutions, and how to connect and analyze sensor data with user data are all KNOW-HOW. In the field of IoT sensors, foreign mainstream manufacturers currently do relatively little because they tend to focus on mass production to reduce costs, while IoT sensors require customization and are produced in small batches. There are still many opportunities in this area.” Wang Yongtao pointed out.
As for the data architecture on the cloud, the corresponding requirements are also evolving. Chen Peng mentioned that once devices are connected to the cloud, all device data and user data will play an important value in the future, and AbleCloud’s big data analysis platform can play a role. On one hand, manufacturers can analyze the status of devices and user behavior based on this massive data, while also profiling users to lay the foundation for subsequent precise push and marketing. On the other hand, only through the mining and processing of massive data can the true “intelligence” of devices be realized, rather than just simple remote control.
“To achieve these services, it also depends on the hybrid data storage services provided by the AbleCloud cloud platform, such as SQL and NoSQL, while also supporting time series. The data storage service is universal, and different manufacturers’ different needs can be met by customizing different data sets and data points on our console platform. Different applications have different definitions of data set formats, which naturally meets different needs.” Chen Peng concluded.
The second season of Smart Product Circle’s “Finding Explosive Products” is currently in the intense final stage. Click “Read Original” to enter the official website and vote for your favorite products!
