Part of the  

Chip Design Magazine


About  |  Contact



Experts Share Unique Challenges in Wearable Designs

By John Blyler, Chief Content Officer

Wearable devices will add a new twist to traditional embedded designs according to experts from ARM, Freescale, HillCrest Labs, STMicr, Imec and Koinix.

Wearable technology design presents challenges different from other embedded markets. To understand these challenges, “System Design Engineering” talked with James Bruce, Director Mobile Solutions for ARM; Mike Stanley, Systems Engineer at Freescale; Daniel Chaitow, Marketing Communications Manager at Hillcrest Labs; Jay Esfandyari, Director of Global Product Marketing at STMicroelectronics; Siebren Schaafsma, Team Leader at Holst Centre and Imec, and; Thea Rejman, Financial Analyst at Kionix, Inc. What follows is a portion of that conversation. – JB

System Design Engineering: What unique technical challenges are designers facing in the wearable smart connected market – as opposed to other markets?

Bruce: The big challenge for wearable designers is that the use cases are still very new.  There is a lot of innovation and diversity taking place. People are trying out many different operational scenarios. Designers need low power processors that are right for these evolving workloads. One of the benefits is that there is a strong ecosystem with a large number of system-on-chip (SoC) available to developers to create initial wearable solutions. Once they have taken the initial designs to market, they can stay with it, use a different SOC, or even customize it.

Another key consideration for designers is improving quality of sensor data integrated into the device, which have traditionally been not in the domain of digital designers. Traditional digital designers need to worry about the analog portions of a wearable device, i.e., accelerometers, gyros, humidity sensors, etc as there are several choices of sensor fusion solutions for low power application processors available off-the-shelf.

Stanley:  Managing power consumption and communications are currently the two largest hurdles for wearable hardware. Looking ahead to truly wearable sensors that can be embedded into clothing, athletic equipments, name badges, etc., means that every component in the system must become even smaller and thinner.  This drives the trend of consolidating multiple sensors into one package, and then shrinking that combo sensor even more. Chip scale packaging will be replacing QFN and LGA for many applications.  This requires close cooperation across all disciplines as the traditional package disappears.

Low power, wireless communications and small form factor are keys to driving IoT applications. In many ways, these are closely related to some wearable applications in that they use similar sensors and common software libraries for communication, data abstraction and signature recognition.

Chaitow: The challenges of timing, power, availability, and security are similar to but different from the typical embedded design problem.  Perhaps the main difference is that, instead of a single circuit board or chip to worry about, the designer must consider the whole system.  This adds network and systems engineering problems to the job of a typical embedded design engineer.  It’s one thing to optimize the power and security of a single device; it’s totally different to do that across a various set of devices in a variable network.  Then add in the fact that the number of devices or versions of the devices might change in a given network over time and the design problem gets even bigger.

An additional unique challenge is the need for calibration of sensors. MEMS sensors used in commercial products have variable performance. This variable performance is true both at the point of manufacture and over the lifetime of the product, as each sensor reacts differently to changes in environmental factors such as temperature, voltage, interference, and sensor aging, to name just a few. These variations make calibration essential for sensor-based products.

Esfandyari: Wearable-device requirements are currently driving significant changes in the MEMS industry. They are driving the development of even smaller components with even lower power consumption but with more embedded features. To satisfy these needs, sensor manufacturers are creating highly integrated devices with multiple sensors (e.g., accelerometer, gyroscope and magnetometer) embedded in a single package.

Finally, from a design perspective, wearable-device manufacturers must be very careful with the appearance of their product because many people are conscious of their appearance and would prefer not to wear accessories that make them look strange. The challenge is to make wearable technology “invisible” to the final user and the external world.

Various implementations of systems that moni-tor activity of the human body. (Courtesy of Imec)

Siebren: An important part of wearable technology will be in the design of body area networks (BANs) – a collection of miniature sensor and actuator nodes. Such devices will require innovative solutions to remove the critical technological obstacles such as shrinking form factors that require new integration and packaging technology. Battery capacities will need to be extended. Indeed, the energy consumption of all building blocks will need to be drastically reduced to allow energy autonomy. System design will have to focus on overall system power consumption where trade-offs have to be made between security, privacy, precision, availability and storage of the data. For example a high power streaming mode over radio of high resolution, medical grade ECG data in case of an emergency compared to average heart rate monitoring once a minute in the low power mode.

Rejman: Stringent power requirements drive the majority of sensor applications in the wearables market. Therefore, it is essential that both the sensors and the software are low power.  Designers should look for a sensor fusion solution that offers embedded power management functionality to help manage sensor interaction and data processing with minimal overhead, resulting in lower power and better performance.

System Design Engineering: Thank you.

Tags: , , , , , , , , , , , , , , , , , , ,

Leave a Reply