Data collection represents the point where the Internet of Things meets the real world. Innovative sensors based on microelectromechanical systems and new optical sensors are pushing back the barriers to ubiquitous IoT deployment. Andrew Pockson, divisional marketing manager, Anglia, tells us more
The Internet of Things (IoT) is driving rapid demand growth for sensors in areas such as asset-tracking systems, smart grids and building automation. According to market research, building automation will generate the largest volumes for microelectromechanical systems and other types of sensors in the industrial IoT market, with asset tracking the second-largest opportunity, driving demand for large volumes of MEMS accelerometers and pressure sensors. Smart grid opportunities include inclinometers to monitor high-voltage power lines as well as accelerometers and flow sensors in smart meters. Environmental monitoring is another key area.
Vendors are responding to demand from the building automation sector with new sensors that put multi-sensor capability into energy-efficient, ultra-low-power packages. For example, the Humidicon range from Honeywell integrates humidity and temperature sensors in a single package. This not only enables a single sensor to do the job of two, but also provides in-built temperature compensation for relative humidity measurement. True compensated digital outputs help to free up space by eliminating the need for signal conditioning components. I2C and SPI outputs are provided to simplify integration to IoT node processors. In addition, low supply voltage (2.3V DC) and low power consumption (1μA in sleep mode) – allows use in low energy and wireless-compatible IoT applications.
Another example of innovations in IoT data collection is Honeywell’s new Amplified Basic Pressure (ABP) sensor, one of the smallest such devices in the industry. Small size and low power requirements, along with as sleep mode and temperature output options, make it a versatile choice for industrial IoT applications.
The advantages of combining multiple sensors in a single package extend into MEMS sensors such as accelerometers and gyro sensors. ST’s MEMS gyroscopes, as well as the accelerometers, magnetometers and iNEMO inertial modules, enable innovative products and apps for industrial solutions including container tracking or emergency-stop mechanisms.
In addition to enabling advanced product miniaturisation, ST’s L3GD20H 3-axis gyroscope also helps conserve battery power by drawing 25 per cent lower current than the previous generation of devices. The sensor’s output has 60 per cent lower noise, which can help to simplify hardware and software design and speed up communication with host systems in IoT nodes.
Instead of combining multiple sensors into a single package, there are other ways of reducing the part-count for MEMS devices. Sometimes it is possible to address the issue laterally, by using alternative underlying technology. An example of this approach is the use of anisotropic magneto-resistance (AMR) sensor for magnetic field detection in three-dimensional space – typically applied to open/close detection in refrigerators and doors, and rotation detection in smart meters. Traditionally this requires the use of multiple magnetic sensors in combination, but the new sensor combining AMR element design technology and circuit design technology from Analog Devices and Murata makes 3D magnetic field detection possible using a single device. The result: products that are more compact and lower in cost.
Innovations in vision
Looking to the future, new optical sensors for vision systems will facilitate applications like intelligent motion sensing, people counting, vehicle detection, and face detection deployed both indoors and out. Technology pioneered by the Swiss Centre for Electronics and Microtechnology has been developed to bring robust, low-cost vision to a wide range of sensing applications in industrial and consumer products.
The sensors identify features within bright light, dark shadow and reflective surface areas. Images are correctly exposed right from the first capture, even when light levels change suddenly. Real-time edge detection and image management are performed in the sensor itself, and inexpensive processors are added to create highly cost-effective, low-power OEM sensing solutions. As a result, the SNAP-S2 image sensor is able to add cost-effective visual sensing capability to a wide range of IoT devices. The sensor is particularly suitable for applications where image consistency and robust feature-detection (Figure 2) are critical. It enables flexible deployment, low system cost and rapid time-to-market – all key requirements when building vision sensing capability into internet linked products.
Whether for mechanical, environmental or optical sensing, sensor fusion takes integration to a higher level, for example combining 3-axis outputs of accelerometer, magnetometer and gyroscope to determine orientation and motion in a three-dimensional space. Sensor hubs perform the real-time algorithms and sensor monitoring independently of the host CPU. They reduce demand on the host processor, lower power consumption and simplify system architecture.
Analog Devices’ iSensor MEMS inertial measurement units are highly integrated, multi-axis solutions that combine gyroscopes, accelerometers, magnetometers, pressure sensors, and additional technology for multiple degrees of freedom applications under dynamic conditions.
Another example is the KX23H sensor hub from Rohm’s dedicated MEMS sensor subsidiary Kionix. Built on a 32-bit ARM Cortex-M0 core running at 32MHz, the KX23H offloads real-time sensor fusion algorithms from a system’s main applications processor. Also on-chip is a high-performance accelerometer, offering 16-bit resolution with a 256-byte FIFO/FILO buffer and very low power consumption down to 1uA. In sleep mode, the ARM core consumes only 2.5uA. It consumes 1.5mA at 32kHz and just 6mA at full speed. The result is a great deal of functionality in a small 3mm x 3mm x 0.9mm package.
Hubs show the way forward
Sensor hubs like these offer comprehensive suites of software including device drivers and fusion libraries optimised for the built-in CPU core, together with sensor algorithms for specific IoT applications. Software and simple programmability allow OEMS to test features and optimise the software quickly, while testing their device in the real world. Stable designs with programmable hubs like the KX23H provide a clear path to custom silicon, with all the cost and power benefits that brings.
With reduced cost, size and power requirements sensors are connecting more of the real world to the Internet of Things.