Skip navigation
Interview: Movea’s Tim Kelliher Discusses Sensor Fusion Trends

Interview: Movea’s Tim Kelliher Discusses Sensor Fusion Trends

Movea has worked with a number of companies such as STmicroelectronics to build sensor fusion platforms (see “Kit Simplifies Sensor Hub Evaluation”). I talked with Tim Kelliher, Customer Solutions Architect at Movea, to find out more about the company and what trends they see coming down the pike.

Wong: Can you tell us a little about Movea?

Kelliher: Movea is the leading provider of data fusion and motion processing firmware, software, and IP for the consumer electronics industry. Our expertise enables customers and partners to unlock the potential for smart applications derived from sensors for the mobile, sports and interactive TV markets while reducing the risk, cost, and time-to-market for delivering compelling new features that differentiate their products and deliver more end-user value.

Since its inception in 2007, Movea has quickly become global with subsidiaries in the United States (San Francisco) and South Korea (Seoul). We also work with technology and manufacturing partners and distributors around the world.

Wong: What type of sensor integration does Movea do?

Kelliher: We turn sensor data into meaningful personal information.  Movea’s SmartMotion technology works with the expanding number of sensors in every smartphone, wearable or other connected device to analyze sensor signals, translating raw sensor data into pertinent information, be it on the device in a sensor hub or off the device as a cloud-based service or a combination of both. We lead the market delivering very low power yet high performance services such as user location, activity monitoring and context awareness, enabling always-on features to be integrated into mobile and wearables.

Wong: Mobile and wearables are very hot topics these days, can you elaborate on how do you position yourself in these markets?

Kelliher: Movea is a thought leader and provider of vital technology in a complex ecosystem for the mobile and wearable markets.  As such, we license our technology either directly to OEMs such as Babolat, Orange and SMK who design motion-enabled intelligent products, or indirectly through licensed semiconductor partners such as STMicroelectronics, Nuvoton and Microchip.  Additionally, we work with other industry leaders such as TI, Atmel, Cadence, Freescale and many others for reference design platforms. This year we announced the most accurate and energy efficient multisport wearable reference design in partnership with TI at CES, as well as the most advanced and low-power sensor hub evaluation kit in partnership with STMicroelectronics at MWC.

Movea sees the wearable and smart devices markets growing exponentially in the near future, as witnessed by the overwhelming buzz at CES this year. To realize the potential, the ecosystem needs to work together to enable a higher level of integration. With this perspective in mind, Movea is sponsoring the Beyond Fusion Conference on June 23 in San Francisco bringing together industry leaders to address the what and how of context awareness for smart devices.

Wong: We hear a lot of people saying that the wearable market growth will be phenomenal; however, consumer adoption is still in the early days and many think the jury is still out as to whether or not it will really take off. What will it take to realize widespread adoption of wearables?

Kelliher: The wearables market is at an inflexion point, moving from the niche geek-gadget adopter to the mass consumer market. To bridge this chasm, the wearable market needs to deliver services that consumers deem valuable without being intrusive into their daily lives. To achieve this, battery life and feature set will need to expand.  To that end, Movea provides the broadest feature set offering on the market including activity monitoring, sleep analysis with automatic detection, and running performance analysis, without draining the battery. We believe that activity monitoring is only the first step in unlocking the full potential of wearables. In the future, we envision them as companion for better living, becoming  an extension of our senses and providing smart recommendations beyond the simple reporting of activities. We expect wearables to evolve, developing features including:

  • Bio monitoring with inclusion of bio sensors such as temperature and heart rate monitor
  • Environment sensing: ambient light, noise, location, pollution
  • Interconnectivity between all connected devices, at home and on-the-go to better sense the user’s context

There is a breadth of features offered today including sensors on the device and in the cloud (Fig. 1).

Figure 1. This chart highlights the breadth of features offered today including sensors on the device and in the cloud.

Wong: What do you see as the next big trend for smart devices?

Kelliher: Given the sensors available today, smart devices are only beginning to scratch the surface of what’s possible in terms of applications. The trend today is “always-on” services such as contextual awareness and indoor location, with sensors and functions running all the time.  Until very recently, power-depleting sensors and signal processing were huge roadblocks to always-on sensor-based services. Since the processing of all these functions was made on the power hungry processor, their number was limited due to the lack of quality signal processing and the amount of power consumed by the application processor draining the battery.  Movea was the first to announce solutions that reduced the average power consumption from ~100 mW to below 5 mW with a sensor hub architecture. At the same time, Movea has expanded the feature set available as an always on-service. Figure 2 indicates that Movea’s mobile offering for features such as activity monitoring, pedestrian monitoring and context awareness capabilities has best-in-class performance without compromising ultra-low power consumption.

Figure 2. Movea’s mobile offering features services like activity monitoring, pedestrian monitoring and context awareness capabilities.

Wong: Siri and Google now have given us a glimpse of your smartphone turning into a personal assistant.  What do you see as the future of the smartphone as a personal assistant?

Kelliher: Today, the smartphone has the ability to recognize your voice, understand basic questions and track your daily activity. Combining those capabilities with the ability to access your email and calendar, the smartphone becomes a very basic personal assistant and activity tracker. In the near future, inputs will be gathered  from several sources including, but not limited to, motion sensors,  audio, WiFi and near field beacons to determine where you are and what you are doing. By combining those inputs with existing services, your smartphone will be a true personal assistant and coach that will be able to understand and anticipate your needs, helping you get to your meetings on time, adjusting your device settings as appropriate and even suggesting when to take a break and go for a walk.

Current smartphone platform architectures provide several processors for each of the subsystems as shown in the diagram below. Each of those subsystems (audio, image, touchscreen, WiFi, GPS, etc.) in turn is capable of providing a portion of the information required to determine the user context such as indoors vs. outdoors, in a car vs. train, sitting at desk, etc. Creating a platform that combines the information from each of the subsystems should help provide a complete picture of a user’s environment, but to avoid draining the battery, a distributed sensor approach is necessary. Such architecture allows the gathering and processing of information from different subsystems at all times without going through the application processor, the largest power hungry component.

Wong: Most people would love to have a reliable indoor navigation application so they’ll never again get lost in places such as a shopping mall or convention center. When will we see this on our smartphones?

Kelliher: Indoor navigation is achieved through a combination of several technologies including inertial dead reckoning, WiFi triangulation, near field beacons (Audio or RF), maps and context. Inertial dead reckoning is a relative displacement technology and can determine how far from a given point you have travelled but not the absolute location that you started. WiFi triangulation and near field beacons provide the initial absolute location. There are uncertainty errors in each of the technologies that can be compensated for with the others. Much like the contextual awareness service, a reliable indoor location service requires a distributed sensor approach wherein there is cooperation among the subsystems to determine when to activate a given subsystem. For instance, the smartphone would only turn on WiFi location, and near field beacons when it has determined that it has gone indoors, or vice versa, disabling those subsystems and turning GPS on when it has detected that it is outdoors. There are many industry working groups tackling this problem today so we expect to see integrated, accurate solutions shortly.

Wong: So what sensors does Movea need to provide navigation and contextual information?

Kelliher: Movea’s pedestrian navigation solution relies on input from accelerometers, gyroscopes and magnetometers to be able to calculate the user’s trajectory and heading, a barometric pressure sensor can also be utilized to determine altitude, such as floor level.  

Basic contextual awareness features such as activity classification to detect whether the user is at rest, walking or running only requires an accelerometer. For advanced features such as detecting the user’s environment, we leverage additional sensors, such as the microphone, to help confirm the accuracy of our findings based on motion. For example, detection of transportation mode (bus, car, train, etc.) can be done with characterizing motion for each of these use cases. By adding audio processing that’s able to recognize specific train or car sounds, the level of confidence increases to understand the user’s context.

Wong: So how does Movea obtain navigation and contextual information from these sensors? Can you give an example?

Kelliher: Raw sensor data is sent to the sensor hub, depending on the architecture. Movea’s patented technology then filters and processes data to transform it into intelligible information that can then be integrated into an end-user application for Android or Windows. For example, let’s take a simple function such as step counting, which you can find in any fitness app today and part of Android 4.4 specifications. Accelerometer data is sent to the processor, then gets filtered, processed, analyzed, and from there, signal patterns are identified as step pattern signals. The number of steps recognized is then leveraged into an Android on Win8 application thanks to an API.

Wong: Sounds very exciting, what is next for Movea in terms of technology developments and partnerships to respond to these trends?

Kelliher: Movea is paving the way to making these always-on sensing and processing features a reality for the smart devices of tomorrow by extending our expertise to additional sensors beyond inertial and magnetic sensors – for example, adding bio sensors to our roadmap. We’re also partnering with industry leaders in adjacent technologies to be able to bring the pieces of the context awareness puzzle together, leading the integration of new data sources to provide always-on context aware services that will simplify users’ lives. Stay tuned for upcoming announcements!

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.